Lumbar Radiofrequency Ablation: Bridging Physics and Clinical Outcomes
Lumbar Radiofrequency Ablation: Physics, Anatomy, Technique, and Evidence
Pain Spa | Dr M. Krishna | Specialist Interventional Pain Management
Introduction and Overview
Lumbar radiofrequency ablation (RFA), also termed radiofrequency denervation or neurotomy, is an interventional procedure that uses thermal energy to ablate the lumbar medial branch nerves that innervate the facet joints, providing relief from chronic facetogenic low back pain. The procedure targets the medial branches of the dorsal rami at the junction of the superior articular and transverse processes — structures that are anatomically constant and fluoroscopically accessible.
Indications: Lumbar RFA is indicated for intractable facet joint pain lasting at least 3 months after failure of conservative measures such as physical therapy, manual manipulation, NSAIDs and acetaminophen. The procedure is typically preceded by diagnostic medial branch nerve blocks (MBBs) with local anaesthetic, with reassessment of pain reduction and functional improvement. NICE guidelines suggest considering RFA after conventional management has not worked in individuals with injection-confirmed facetogenic pain.
Evidence status: The evidence is mixed. A 2015 Cochrane review found moderate-quality evidence for greater short-term pain relief compared with placebo. The 2017 MINT trial found no clinically important differences for facet joint pain. The 2022 VA/DoD guidelines note moderate-quality evidence for pain improvement at 6, 12 and 36 months. A 2025 multicentre trial comparing RFA to cryoablation and endoscopic denervation found all three effective, with RFA providing the most consistent long-term relief.
RF Physics: Mechanism of Heat Generation
RFA uses high-frequency alternating electrical current in the radiofrequency range of 300–500 kHz (300,000–500,000 Hz) — in the radio wave portion of the electromagnetic spectrum. Critically, the mechanism of heating is not electromagnetic wave radiation, as with microwaves or lasers, but resistive or ohmic heating of tissue acting as an electrical conductor.
Resistive/Ionic Heating — The Core Mechanism
At approximately 400 kHz, the wavelength is around 600 metres — far too long for tissue absorption as electromagnetic radiation. The mechanism is purely electrical.
Alternating current flows through tissue: The RF generator delivers alternating current between the small active electrode, which is the needle tip, and the large dispersive electrode, which is the grounding pad. Tissue completes the electrical circuit.
Ionic agitation: The alternating electric field causes charged ions such as Na⁺, K⁺ and Cl⁻ to oscillate back and forth at the current’s frequency. This rapid oscillation creates frictional heating as ions collide with surrounding molecules.
Tissue is the heat source, not the electrode: The electrode does not generate heat. The tissue surrounding the electrode becomes the primary heat source through resistive losses. The temperature sensor at the tip measures tissue temperature — not the electrode’s own temperature.
Why 300–500 kHz?
Above the neuromuscular stimulation threshold: At lower frequencies, alternating current causes depolarisation of nerve and muscle cells, resulting in pain and contraction. Above approximately 10 kHz, the current alternates too rapidly for cell membranes to depolarise — it passes through tissue producing only heat.
Below the microwave range: At microwave frequencies, the mechanism shifts to dielectric heating of water molecules through polar molecule rotation — a fundamentally different and less controllable process for focal ablation.
Two Sequential Phases of Heating
Resistive heating phase: Immediate heating of a 1–2 mm rim of tissue in direct contact with the electrode. Approximately 90% of delivered power is absorbed within the first 1–1.5 mm from the electrode surface. This is the zone of direct ionic frictional heating.
Conductive heating phase: The resistively heated zone acts as a secondary heat source, and thermal energy passively conducts outward. This phase is time-dependent — longer ablation durations allow more conductive spread, which is why lesion size increases with duration.
The Bio-Heat Equation
Coagulation necrosis = (heat generated × local tissue interactions) − heat lost
In practical terms, useful lesioning depends on current density around the active tip, tissue impedance, distance from the electrode, duration of energy delivery, local heat conduction, and convective heat loss from blood flow. This explains why lesion size is not determined by generator temperature alone.
Heat is lost through conduction to normothermic surrounding tissue and convection via blood flow in nearby vessels — explaining the “heat sink” effect near large vessels that limits lesion size.
| Temperature | Tissue Effect |
|---|---|
| 42–45°C | Onset of irreversible cellular damage with prolonged exposure |
| 46–49°C | Irreversible damage for most mammalian tissues |
| 50°C | Threshold for coagulation necrosis |
| 80°C | Standard target for conventional pain RFA |
| >100°C | Tissue desiccation, charring and impedance rise — reduces effectiveness |
| >110°C | Tissue acts as an electrical insulator; ablation ceases |
| Modality | Frequency | Heating Mechanism |
|---|---|---|
| RFA | 300–500 kHz | Resistive/ionic friction: electrical current through tissue |
| Microwave ablation | 900–2500 MHz | Dielectric heating: polar water molecule oscillation in electromagnetic field |
| Laser ablation | ~1064 nm infrared | Photon-to-kinetic energy transfer |
| HIFU | Ultrasound | Mechanical vibration and cavitation |
Functional Anatomy
Medial Branch at L1–L4
At L1–L4, the dorsal ramus divides into medial, lateral and intermediate branches within the intertransverse ligaments. The medial branch runs across the superior portion of the transverse process, passes under the mamillo-accessory ligament at the junction of the superior articular process and the root of the transverse process, and then courses onto the lamina — innervating the facet joints at that level and the level below.
Why the Dorsal Ramus is Targeted at L5
At L5, the anatomy is fundamentally different. The L5 dorsal ramus is much longer than other lumbar dorsal rami and does not divide until a more distal location. It runs in a groove between the superior articular process of S1 and the sacral ala. At the caudal edge of the S1 superior articular process, the ramus divides into medial and lateral branches, with the medial branch supplying the L5–S1 facet joint.
Clinical rationale: Because the L5 dorsal ramus travels a longer course before branching, and is accessible in the groove at the S1 superior articular process and sacral ala, it is more practical to target the entire dorsal ramus at this level. Its diameter is approximately 0.5 mm transverse — making precise targeting critical.
Anatomical variability: Motor stimulation responses are typically less prominent at L5, possibly due to greater adipose tissue, multifidus atrophy, or because it may actually be the medial branch, rather than the dorsal ramus, being denervated in some patients.
The Mamillo-Accessory Ligament (MAL)
The mamillo-accessory ligament (MAL) is a fibrous band bridging the mamillary process, which is the rounded eminence on the posterior aspect of the superior articular process, and the accessory process, which is the small projection at the transverse process–pedicle junction. It forms an osseofibrous tunnel through which the medial branch passes. It is thought to represent remnants of transversospinal elements.
Fixes the nerve to bone: The MAL tethers the medial branch to a predictable location — the “eye” on oblique fluoroscopy. This anatomical constancy is what makes percutaneous medial branch blocks and RFA reliably accurate.
Nerve occupies only approximately 3% of the tunnel: The nerve is surrounded by adipose tissue, providing a protective cushion and some room to move.
Ossification: The MAL can ossify into a bony foramen, called the mamillo-accessory foramen. Recent data found overall ossification prevalence of 72.73% across all lumbar levels, beginning at age 30–45 and increasing with age. It is most common at L5, reported as 26% on the left and 13.5% on the right in Maigne et al., and more prevalent in females, with 76.92% at L5 left. An ossified MAL may prevent direct needle access and reduce lesion effectiveness — particularly relevant at L5 where ossification is most common.
Diagnostic Medial Branch Blocks
MBBs vs. Intra-Articular Injections
Medial branch blocks serve as prognostic tests — not therapeutic interventions. Evidence shows that MBBs provide better predictive value for RFA success than intra-articular facet injections. Both techniques show better predictive value compared with sham blocks.
Technical Specifications
Volume: Less than 0.5 mL of local anaesthetic improves specificity and reduces spread to adjacent structures.
Agents: Lidocaine, bupivacaine or ropivacaine may be used. The MINT trial used 0.5 mL of 2% lidocaine.
Timing: A block is considered positive if the patient reports adequate pain reduction within 30–90 minutes.
Avoid steroids: Steroids may confound diagnostic results. Exceptions include patients at risk for RFA complications, such as young athletes, those on anticoagulation, or patients with cardiac devices, where steroids may provide intermediate-term relief.
Pain Relief Threshold
At least 50% pain relief is the most widely accepted cutoff, but this remains debated.
Higher thresholds, such as ≥80%, are associated with better RFA outcomes: 84% RFA success with ≥80% block relief compared with 56% success with 50–80% relief.
Patients selected with ≥80% relief on dual comparative blocks showed 63.2% success at 6–12 months and 65.6% at 12–24 months.
Functional improvement should be considered alongside pain reduction.
Single vs. Dual Blocks
| Approach | Advantages | Disadvantages |
|---|---|---|
| Single block (ASPN consensus) | Balances false-positives and false-negatives; allows personalised medicine; Grade C recommendation | More false-positives than dual approach |
| Dual blocks (ASIPP / SIS) | Minimises placebo effect; higher RFA success rates | Excludes patients who would benefit; higher cost |
| No blocks | Highest number of positive RFA responses; lowest cost per successful treatment | Includes all placebo responders; lowest specificity |
False-positives: Injectate spillage, use of sedation, excessive superficial anaesthesia, and resting after the block rather than testing usual painful movements.
False-negatives: Intravenous uptake of anaesthetic, failure to anaesthetise the target nerve, aberrant anatomy, procedure-related pain, and opioid-induced hyperalgesia.
Lesion Size — Determinants and Dimensions
Key Determinants
Understanding lesion size variables is critical because lumbar medial branches are less than 2 mm in diameter and the L5 dorsal ramus is only 0.5 mm. Three primary factors determine lesion size:
Distance from the cannula’s active tip
RF current density
Duration of current application
Device-Related Factors
Cannula diameter: Perhaps the most efficient method to increase lesion size. Increasing from 22G to 16G at 80°C for 2 minutes increases average lesion width by 58–65%, equivalent to approximately 3–4 mm larger.
Active tip length: Longer tips create larger lesions, extending 1–5 mm distal and proximal to the tip at 80°C for 2 minutes. Lesion length correlates almost perfectly with tip exposure (r² = 0.996).
Temperature: Increasing from 60°C to 90°C for 2 minutes increases lesion width by 108–152%. Temperatures beyond 90°C are not recommended — tissue boiling and charring create high impedance insulation, paradoxically reducing lesion size.
Duration: Approximately 87% of maximal lesion surface area occurs within the first 60–90 seconds. Extending lesion time from 1 to 3 minutes increases lesion size by 23–32% and reduces variability.
Lesion Dimensions by Gauge — Monopolar RF at 80°C, 10 mm Active Tip
| Gauge | Approx. Lesion Width | Lesion Volume (ex vivo, 90s) | Notes |
|---|---|---|---|
| 16G | ~8–10 mm | ~360 mm³ | Largest monopolar lesion; approaches cooled RF |
| 18G | ~7–8 mm | ~169 mm³ | Most common clinical choice |
| 20G | ~6–7 mm | Not specified | 18G preferred clinically |
| 22G | ~5–6 mm baseline | Not specified | Reference standard in studies |
These are ex vivo measurements that likely underestimate in vivo lesion size. Proximity to bone approximately doubles the maximal effective radius compared with muscle-only models — directly relevant when the electrode is positioned against the transverse process or sacral ala.
Active Tip: 5 mm vs. 10 mm
Active tip length primarily affects lesion length rather than width. A 5 mm active tip produces a lesion approximately 6–10 mm long; a 10 mm active tip produces a lesion approximately 11–15 mm long.
Other Technical Modifiers
Proximity to bone: Approximately doubles the maximal effective radius compared with muscle-only models.
Fluid pre-injection: 2% lidocaine increases lesional width; contrast agents such as iohexol increase length.
Adipose tissue: Reduces lesion size compared with muscle due to lower thermal conductivity.
Needle Positioning: Parallel vs. Perpendicular
Why Parallel Placement is Preferred
Parallel needle placement is preferred because the RF lesion extends radially around the active tip in an oblate spheroid shape but does not extend beyond the distal tip. With a perpendicular approach, the nerve must fall within a narrow approximately 2 mm radial window at the tip — if the nerve is even slightly off-target, it escapes the lesion entirely.
Bogduk et al. (1987) demonstrated in egg white and meat models that lesions never extend distal to the electrode tip. Lau et al. (2004) confirmed in cadavers that with perpendicular placement, nerves were situated at the tapering end of the elliptical lesion and were frequently not captured.
Lesion geometry principle: A conventional RF lesion is widest around the active tip but tapers at the distal end. Parallel or obliquely parallel placement increases the length of nerve exposed to the effective thermal field; perpendicular placement leaves the nerve dependent on a very small capture zone.
The Nuance: Obliquely Parallel
Cohen et al. (2020) note that if the electrode were placed in the exact same plane as the nerve, truly parallel placement could minimise ablation because the nerve could run between the electrode and bone without being captured. The optimal trajectory is therefore obliquely parallel — inserted in a medial-cephalad direction so the active tip crosses the neck of the superior articular process where the nerve courses.
Clinical Evidence
Loh et al. studied 323 patients and found near-parallel placement produced longer duration of relief, with a median of 4 months compared with 1.5 months (p=0.02), and lower 1-month pain scores, 3.64 compared with 4.27 (p=0.06). Important caveat: suboptimal perpendicular placement can still produce adequate sensory stimulation at less than 0.5V because the electrical field extends further than the thermal lesion — a positive sensory test does not guarantee nerve capture.
Multi-Tined Electrodes: The Exception
Multi-tined electrodes, such as Trident cannulae, are designed for a perpendicular approach, with tines that deploy outward to create a wider lesion footprint. Deng et al. (2022) compared multi-tined perpendicular placement to conventional monopolar parallel placement in 51 patients and found equivalent pain relief, 52% versus 57% NRS reduction, duration of relief of 8.7 versus 8.4 months, and equivalent disability scores, with the multi-tined approach being faster and requiring less radiation.
RF Modalities: Conventional, Multi-Tined, and Cooled RF
Conventional Monopolar RF
Conventional monopolar RF is the most established technology. It generates heat via high-frequency alternating current, creating an oblate spheroid lesion. Typical settings are 80°C for 90–120 seconds using an 18G cannula with a 10 mm active tip.
Advantages: Extensive evidence base, predictable lesion geometry and well-established technique.
Disadvantages: Technically demanding parallel placement; anatomic variability can lead to nerve non-capture.
Note: A 16G cannula at 80–90°C for 2–3 minutes generates lesion widths approaching cooled RF.
Multi-Tined (Trident) RF
Multi-tined RF deploys three tines that splay outward from the introducer tip, creating a wider lesion footprint that compensates for anatomic variability. It uses a perpendicular approach.
Advantages: Faster procedure time (31.1 vs. 37.6 min, p<0.001), less local anaesthetic use (11.0 vs. 15.8 mL, p<0.001), and less radiation exposure (30.2 vs. 41.5 mGy, p=0.05).
Efficacy: Equivalent to conventional monopolar RF in all pain and disability measures in Deng et al. (2022).
Cooled RF (CRFA)
Cooled RF uses an internally water-cooled electrode, typically 17G with a 4 mm active tip, that prevents tissue charring near the electrode and allows greater current deposition into deeper tissue. The set temperature is 60°C.
Largest lesion volume: Approximately 595 mm³ ex vivo — significantly larger than any conventional monopolar configuration.
Ablation distal to tip: Significant tissue ablation occurs distal to the tip, unlike conventional electrodes.
Lesion shape: More spherical; placement too close to bone distorts the lesion into an incomplete shape.
Safety consideration: Risk of skin burns in thin patients — a case of third-degree burn from thoracic cooled RF was reported in a patient with BMI 21.8.
| Technology | Gauge | Active Tip | Temp | Time | Lesion Volume (ex vivo) |
|---|---|---|---|---|---|
| Conventional monopolar | 20G | 10 mm | 80°C | 90s | ~105 mm³ |
| Conventional monopolar | 18G | 10 mm | 80°C | 90s | ~169 mm³ |
| Conventional monopolar | 16G | 10 mm | 80°C | 90s | ~360 mm³ |
| Protruding electrode (PERF) | 18G | 10 mm | 80°C | 150s | ~283 mm³ |
| Cooled RF | 17G | 4 mm | 60°C | 150s | ~595 mm³ |
Clinical Outcomes Comparison
| Outcome | Conventional Monopolar | Multi-Tined | Cooled RF |
|---|---|---|---|
| ≥50% NRS reduction at 6 months | 64.7% | — | 74.1% |
| NRS reduction at 6 months | −3.4 points | — | −4.3 points |
| NRS reduction at 3 months | −4.0 pts (52%) | −4.3 pts (57%) | — |
| Duration of relief | 8.7 months | 8.4 months | Comparable |
| ODI improvement | Significant | Equivalent | Significant |
| Procedure time | 37.6 min | 31.1 min ✓ | — |
| Radiation dose | 41.5 mGy | 30.2 mGy ✓ | — |
| Local anaesthetic | 15.8 mL | 11.0 mL ✓ | — |
| Needle approach | Parallel | Perpendicular | Parallel or perpendicular |
| Adverse events | None reported | None reported | Skin burn risk in thin patients |
Key evidence: Provenzano et al. (2026) multicentre RCT (n=74): cooled RF met non-inferiority compared with standard RFA (p=0.0069). Shih et al. (2020) meta-analysis: efficacy ranked cooled RF > thermal RF > PRF at 6 months — with no statistically significant differences between techniques. Deng et al. (2022): multi-tined RF offers equivalent outcomes with significant procedural advantages. Technique and patient selection appear more important than electrode technology.
RFA vs. Cryotherapy — Lesion Size and Mechanism
Cryotherapy Ice Ball Dimensions
Cryotherapy ice balls are substantially larger than conventional RF lesions, but the effective ablation zone within the ice ball is much smaller than the visible ice ball — a critical clinical distinction.
Said et al. (2023): Probe gauge is the dominant factor in ice ball size. Changing from 18G to 14G increased ice ball width by up to 70%, length by up to 113%, and volume by up to 512%. A 17G (2.4 mm) probe produces an approximately 30–31 mm ice ball diameter.
Ice Ball ≠ Ablation Zone: The Critical Distinction
The visible ice ball overestimates the effective ablation zone because ice forms at temperatures between 0°C and −20°C, but only temperatures below −20°C induce Wallerian degeneration, which is the desired nerve injury for cryoneurolysis.
The −20°C isotherm sits approximately 1 cm inside the leading edge of the visible ice ball.
In a porcine lung model with a 2.4 mm probe, the ice ball diameter was 31 mm, but the −20°C isotherm was only 23 mm in diameter.
Ilfeld et al. (2026) reported that tissue temperatures adjacent to the cryoprobe shaft frequently did not reach −20°C even after 3 minutes, with a 71% failure rate, improving to 42% failure at 5 minutes.
| Parameter | Conventional RFA (18G, 80°C, 10mm tip) | Cryoneurolysis (18G probe) | Cryoneurolysis (14G probe) |
|---|---|---|---|
| Visible lesion/ice ball width | ~7–8 mm | Much larger (~15–20 mm estimated) | Up to 70% wider than 18G |
| Effective ablation zone | ~7–8 mm; entire lesion ablative | Smaller than ice ball; ≤−20°C zone | Larger, but still smaller than ice ball |
| Lesion shape | Oblate spheroid around active tip | Near-spherical around probe | Same shape, larger |
| Extends beyond tip? | No | Yes | Yes |
Needle/Probe Positioning for Cryotherapy
Unlike RF where parallel placement is critical, cryoneurolysis allows a perpendicular approach because the ice ball forms a large, roughly spherical zone that extends in all directions — including distally beyond the tip. The nerve simply needs to be within the ice ball radius regardless of probe orientation.
| Feature | Conventional RF | Cryoneurolysis |
|---|---|---|
| Optimal approach | Parallel to nerve | Perpendicular or oblique |
| Lesion shape | Oblate spheroid; radial only | Near-spherical; all directions |
| Extends beyond tip? | No | Yes |
| Placement precision needed | Very high | Moderate, within ~3 mm |
| Real-time monitoring | Sensory stimulation | Ultrasound visualisation of ice ball |
| Forgiveness of misplacement | Low | High due to large ablation zone |
Clinical outcomes: Head-to-head clinical trials comparing RFA and cryoablation for lumbar facet pain show comparable outcomes at up to 24 months. One pilot study involving 30 patients suggested cryoneurolysis may provide superior outcomes at 12 months, although this requires confirmation in larger trials.
Nerve Injury Classification and Regeneration
Sunderland Classification
| Feature | Cryoneurolysis (2nd Degree) | Conventional RFA (3rd Degree) |
|---|---|---|
| Axon | Destroyed | Destroyed |
| Myelin | Destroyed | Destroyed |
| Endoneurium | Preserved ✓ | Destroyed |
| Perineurium | Preserved ✓ | Preserved ✓ |
| Epineurium | Preserved ✓ | Preserved ✓ |
| Fascicular arrangement | Preserved ✓ | Preserved ✓ |
| Regeneration | Orderly; guided by intact endoneurial tubes | Guided by intact perineurium/epineurium |
| Neuroma risk | Very low | Exceedingly rare |
| Time to recovery (animal models) | ~8 weeks | Variable, less predictable |
| Post-procedure neuritis | Less common | More common, approximately 7–13% |
Why RFA Rarely Causes Neuromas
Choi et al. (2016) described conventional RFA as producing a Sunderland third-degree injury — destruction of the myelin, axon and endoneurium, without disruption of the fascicular arrangement, perineurium or epineurium. Because the perineurium remains intact, regenerating axons are confined within the fascicle, preventing the disorganised extrafascicular sprouting that leads to neuroma formation.
Schmidt et al. (2021) found only one case report in the literature describing neuroma formation following lumbar medial branch RFA. Neuroma formation, which is seen in fourth- and fifth-degree injuries where the perineurium or entire nerve trunk is disrupted, is therefore exceedingly rare with RFA. The preserved perineurium also allows predictable nerve regeneration — explaining why pain recurs and repeat procedures are eventually needed.
Grounding Pad: Placement and Safety
Standard Placement
In monopolar RFA, current flows from the small-area electrode tip through the body to the large-area grounding pad. The massive difference in cross-sectional area concentrates energy at the tip — the grounding pad’s large surface area ensures current density there remains low enough to avoid thermal injury.
Lumbar/SIJ RFA: The grounding pad is placed on the ipsilateral lower extremity, usually the anterior thigh.
Distance: At least 25 cm from the electrode; ASA recommends at least 15 cm from pacing leads in patients with cardiac implantable electronic devices.
Does Distance Matter?
Lesion size: Concordant dispersive patch placement produced approximately 20% greater lesion depth, 7.4 vs. 6.1 mm (p=0.001), in an in vivo swine study by Venkateswaran et al. (2025). However, this was at cardiac ablation current densities, so the clinical significance for medial branch neurotomy is uncertain.
Burn risk: Goldberg et al. (2000) found temperature elevations at the grounding pad in 68% of trials, with second-degree burns at temperatures greater than 47°C and third-degree burns at 52°C or above. Grounding pad heating depends on distance, surface area and orientation. Closer placement causes more heating, larger surface area causes less heating, and placing the longest edge facing the electrode reduces heating.
Ipsilateral vs. Contralateral Placement
Nath et al. (1996) found dispersive electrode position had no significant effect on electrical parameters or tip temperature. For standard lumbar RFA without implanted devices, ipsilateral versus contralateral placement is clinically interchangeable. When implanted devices are present, laterality should be chosen to route the current pathway away from the device. For example, for a right-sided gluteal spinal cord stimulator IPG undergoing left lumbar RFA, ipsilateral left leg placement routes current away from the device.
Practical Recommendations
Use a large electrical dispersive pad constructed with conductive metal and adhesive polymer gel.
Position the pad with the longest side facing the RF electrode.
For high-risk procedures, consider dual grounding pads to increase total surface area.
Ensure good skin contact with clean, dry and well-prepared skin.
RFA in Patients with Implanted Devices
Spinal Cord Stimulators (SCS)
Primary concerns: Electromagnetic interference may cause overstimulation, unintended tissue damage, device damage, or permanent loss of therapy.
Preoperative: Determine device manufacturer and IPG/lead location. Place the device in surgery mode if available — this releases a small protective current preventing the lead from becoming a ground for the RFA. If there is no surgery mode, turn the device off. Informed consent should include the risk of IPG damage requiring reimplantation.
Intraoperative: Bipolar RFA is preferred because the electromagnetic field is confined to tissue between the two electrodes. If monopolar is used, place the grounding pad so the needle-to-pad distance is less than the needle-to-IPG/lead distance. Use minimal sedation; patients should report any abnormal sensations.
| Manufacturer | Key Recommendation |
|---|---|
| Abbott | Do NOT use monopolar; use surgery mode; bipolar only |
| Boston Scientific | Bipolar recommended; do NOT use monopolar; turn off stimulation ≥5 min before |
| Medtronic | RFA may cause overstimulation/device damage; route current away from device |
| Nevro | Safety not established for RFA with implanted SCS leads; turn IPG off; use lowest energy |
Cardiac Implantable Electronic Devices: Pacemakers and ICDs
Key anatomical principle: EMI from procedures below the umbilicus, approximately L3–4, is unlikely to cause significant CIED interference, as long as the return pad is also below the umbilicus. Lumbar medial branch RFA is inherently lower risk than cervical RFA for CIED patients.
Pacemaker-dependent patients: Programme to asynchronous mode, such as DOO, AOO or VOO, or place a magnet over the device.
ICD patients: Programme tachyarrhythmia detection off, or place a magnet over the device to suspend antitachycardia function. A magnet suspends tachyarrhythmia detection but has no effect on pacing function.
Intraoperative: Position the grounding pad so the current pathway does not pass near the pulse generator or leads. Have external defibrillator and temporary pacing available. Monitor with at least two separate methods. Use bipolar RFA when possible.
| Manufacturer | Key Recommendation |
|---|---|
| Abbott | Programme tachyarrhythmia therapies off; asynchronous pacing mode; ground plate under buttocks/legs; external defib available |
| Boston Scientific | Electrocautery Protection Mode or asynchronous pacing; bipolar preferred; external defib/pacing available; check thresholds post-procedure |
| Medtronic | Asynchronous pacing via magnet or programming (DOO); route current away from device; monitor with at least two methods |
Bipolar vs. Monopolar: Why Bipolar is Safer
In monopolar RFA, current flows from the needle tip through the entire body to the distant grounding pad, creating a large electromagnetic field that can encompass implanted devices. In bipolar RFA, current flows only between two closely spaced electrodes, confining the electromagnetic field to a small tissue region.
Hanna & Abd-Elsayed (2021) reported no adverse events or device interactions in 33 patients undergoing 71 bipolar RFA treatments with CIEDs. The trade-off is that bipolar RFA produces a different lesion geometry, cylindrical between the two electrodes, and requires precise placement of both needles.
Deep Brain Stimulators (DBS)
Both Abbott and Medtronic state that safety has not been established for RFA in patients with DBS, with specific concern about induced currents causing heating at the lead-electrode site resulting in brain tissue damage. Boston Scientific DBS guidelines allow bipolar or monopolar electrocautery but require probes at least 1 inch from the implanted device. Despite these warnings, safe use of monopolar lumbar RFA has been reported in a patient with DBS and an abdominal wall IPG.
Pulsed RF — Mechanism and Role
Mechanism: Neuromodulation, Not Ablation
Pulsed RF (PRF) delivers short bursts of current, typically 20 ms pulses every 0.5 seconds, with intermittent pauses. This keeps tip temperature at or below 42°C and achieves neuromodulation without thermal destruction. PRF works through fundamentally different biological pathways:
Generates high electric fields, approximately 150 kV/m near the electrode tip, causing microscopic structural changes in cell membranes and organelles without macroscopic thermal damage.
Primarily modulates C-fibre signalling while leaving myelinated fibre conduction intact.
Alters expression of ion channels including Na/K ATPase, HCN and P2X3; neurotransmitters including glutamate and aspartate; inflammatory cytokines including IL-6 and TNF-α; and intracellular signalling cascades including ERK1/2, p38 and JNK.
PRF and EMI — A Nuanced Risk
PRF theoretically produces less EMI due to lower energy and intermittent current. However, Medtronic specifically warns that pulse-modulated ablation systems may pose higher risk for induced ventricular tachyarrhythmias in patients with CIEDs. The pulsatile delivery pattern may be more likely to be sensed as cardiac activity by CIED sensing algorithms, even though total energy is lower. Lower thermal energy does not automatically mean lower EMI risk.
Efficacy vs. Conventional RFA for Facet Joint Pain
Tekin et al.: Conventional RFA maintained improvement at 6 and 12 months; PRF did not maintain improvement beyond the immediate post-procedure period.
Shih et al. (2020) meta-analysis: Ranked efficacy: cooled RF > thermal RF > PRF, with PRF least effective for lumbar facet joint pain.
AAPM guidelines: Thermal RFA produces clinically and statistically significant improvements in pain and function at 3 months compared with pulsed RF, with Level II evidence.
Where PRF Has a Role
DRG application for radicular pain: Strong evidence for modulation without ablation.
Postherpetic neuralgia, trigeminal neuralgia and occipital neuralgia: PRF may avoid motor deficits associated with thermal ablation.
Patients with implanted devices: Consider only if bipolar RFA is unavailable and monopolar risk is deemed unacceptable — with the understanding that efficacy for facet pain will be substantially lower. ASPN guidelines recommend bipolar conventional RFA, not PRF, as the preferred approach for this patient group.
Complications and Safety Profile
The overall safety profile of lumbar RFA is favourable, with complications typically self-limited. Most studies were not designed to evaluate adverse events systematically, so true complication rates remain uncertain.
Localised pain and tenderness: Lasting up to 1–3 weeks, reported in 3.3–8.8% of procedures.
Hypoesthesia or dysesthesia: Usually temporary and affecting the overlying skin.
Transient non-painful paresthesias.
Mild lower limb weakness: Resolves completely.
Rare: Superficial burns, exacerbation of pain, or change in pain characteristics.
Neuroma: Exceedingly rare, with a single case report in the literature described by Schmidt et al. (2021).
Post-procedure management: Injection of steroid through the cannula after ablation but before removal may reduce postprocedural pain and discomfort, with a Grade C recommendation.
Summary: Key Clinical Principles
| Topic | Key Principle |
|---|---|
| RF heating mechanism | Resistive/ionic, not radiation — tissue is the heat source, not the electrode; 300–500 kHz alternating current |
| L5 target | Dorsal ramus, not medial branch — longer course, accessible in groove between S1 SAP and sacral ala; approximately 0.5 mm diameter |
| MAL | Fibro-osseous tunnel fixes medial branch to bone at L1–L4; can ossify, especially at L5, with up to 72.73% overall, potentially interfering with RFA |
| Diagnostic MBBs | MBBs superior to IA injections for RFA selection; ≥50% relief threshold most common; single vs. dual blocks remains controversial |
| Lesion size | Larger gauge = larger lesion: 16G > 18G > 20G > 22G; longer active tip = longer lesion; longer duration = larger, more consistent lesion; bone proximity doubles effective radius |
| Needle positioning | Parallel to nerve for conventional/cooled RF; perpendicular for multi-tined; obliquely parallel is optimal for conventional |
| RF modalities | Conventional, multi-tined and cooled RF produce equivalent clinical outcomes; technique and patient selection matter more than electrode technology |
| RFA vs. cryo | Ice ball much larger than RF lesion, but effective ablation zone below −20°C is approximately 1 cm inside the ice ball edge; cryo allows perpendicular approach |
| Nerve injury | RFA = Sunderland 3rd degree, with endoneurium destroyed and perineurium/epineurium preserved; cryo = 2nd degree; both reversible; neuroma exceedingly rare |
| Grounding pad | Ipsilateral thigh standard; ipsilateral vs. contralateral equivalent without implanted devices; route current away from implanted devices |
| Implanted devices | Bipolar preferred over monopolar; lumbar RFA below umbilicus is lower cardiac device risk; device-specific protocols mandatory |
| Pulsed RF | Neuromodulation, not ablation; ≤42°C; inferior to thermal RFA for facet pain; not necessarily safer for implanted device patients and may increase arrhythmia risk |
Pain Spa Expert Perspective
The clinical success of lumbar radiofrequency denervation depends on three linked domains: correct diagnosis, accurate anatomy, and correct lesion physics. A positive medial branch block is only the starting point. The procedure then depends on understanding the small diameter of the medial branches, the unique anatomy of the L5 dorsal ramus, the tethering effect of the mamillo-accessory ligament, and the limitations of lesion geometry.
In practical terms, careful image-guided placement and adequate lesion creation matter more than simply selecting a particular RF generator. Larger cannulae, appropriate active tip length, adequate duration, safe temperature selection, and obliquely parallel electrode placement all increase the likelihood that the target nerve is captured within the effective thermal lesion. Cooled RF and multi-tined systems may compensate for anatomical variability, but they do not remove the need for precise anatomical understanding.
For patients, the key message is that lumbar RFA is not a general treatment for all back pain. It is a targeted treatment for carefully selected facet-mediated pain, ideally confirmed by diagnostic medial branch blocks and performed using meticulous fluoroscopic technique. When performed appropriately, it can provide meaningful relief while preserving the possibility of repeat treatment as the nerve regenerates.
Dr Krishna has extensive experience in advanced image-guided neuromodulation and radiofrequency techniques, including conventional thermal radiofrequency, pulsed radiofrequency, cooled radiofrequency and targeted peripheral nerve procedures. His approach combines detailed anatomical knowledge, careful diagnostic assessment and precise fluoroscopic or ultrasound-guided technique to improve the likelihood of accurate nerve targeting and meaningful clinical benefit.
Patients with persistent low back pain thought to arise from the facet joints, or those who have been advised to consider radiofrequency denervation, are welcome to contact Pain Spa for a specialist assessment and personalised treatment plan.
Key References
1. Sayed D, et al. ASPN Evidence-Based Clinical Guideline of Interventional Treatments for Low Back Pain. Journal of Pain Research. 2022.
2. Lee DW, et al. LEARN: Best Practice Guidelines from ASPN for Radiofrequency Neurotomy. Journal of Pain Research. 2021.
3. Cohen SP, et al. Consensus Practice Guidelines on Interventions for Lumbar Facet Joint Pain. Regional Anesthesia and Pain Medicine. 2020.
4. Provenzano DA, et al. Non-Inferiority Multicenter RCT Comparing Cooled RFA to Standard RFA for Facetogenic Lumbar Pain. Regional Anesthesia and Pain Medicine. 2026.
5. Deng G, et al. Prospective Within-Subject Comparison of Multi-Tined vs. Conventional Monopolar Cannula for Lumbosacral Facet Joint RFA. Pain Physician. 2022.
6. Shih CL, et al. Comparison of Efficacy Among Different RFA Techniques for Lumbar Facet and SI Joint Pain: Systematic Review and Meta-Analysis. Clinical Neurology and Neurosurgery. 2020.
7. Bogduk N. The anatomy of the lumbar dorsal ramus and its medial branches. 1982; updated in subsequent publications.
8. Choi GS, et al. Nerve injury grades after radiofrequency neurotomy. Korean Journal of Pain. 2016.
9. Schmidt PC, et al. Neuroma formation following lumbar medial branch RFA: narrative review. 2021.
10. Hanna MN, Abd-Elsayed A. Systematic review: bipolar RFA in patients with CIEDs. 2021.
11. Poodendaen C, et al. Prevalence and clinical significance of mamillo-accessory foramen ossification. 2024.
12. Said N, et al. Ex vivo cryoneurolysis probe gauge and ice ball characteristics. 2023.
13. Knezevic NN, et al. Low Back Pain. Lancet. 2021.
14. Maigne JY, et al. Ossification of the mamillo-accessory ligament in 203 lumbar spines. Spine. 2014.