Percutaneous analgesic device enabling both local anesthetic delivery and electrical stimulation (neuromodulation) of peripheral nerves: a pilot feasibility study (case series) Authors: Brian M Ilfeld, Engy T Said, Brian H Park, Sanjay K Sinha, Bryan Leek, Matthew J Meunier, Ian M Foran, Andrew Hurvitz, William T Kent, Alexandra K Schwartz, Baharin Abdullah, Rafael Alkabalan, Nathan Lau, John J Finneran Published: 2026 ABSTRACT AbstractBackgroundA novel device integrating both local anesthetic delivery and peripheral nerve stimulation (PNS) to treat postoperative pain is under development. The device uses a catheter-over-needle design that permits ultrasound-guided percutaneous insertion. An integrated electrode and pulse generator enable PNS for up to 28 days. Such an approach may represent a paradigm shift in postoperative pain management by enabling the delivery of (1) a single-injection peripheral nerve block, (2) a continuous peripheral nerve block, and (3) neuromodulation—all through a single system that can be placed in a timeframe comparable with that of a traditional single-injection nerve block. The current prospective pilot study was executed under a US Food and Drug Investigational Device Exemption to develop insertion and management protocols, as well as assess the feasibility and safety of using the device to treat postoperative pain.MethodsPreoperatively, adults (n=20) undergoing moderate-to-severely painful ambulatory shoulder or foot/ankle surgery had a device (RELAY, Gate Science, Moultonborough, NH) inserted under ultrasound guidance adjacent to the brachial plexus or sciatic nerve, respectively. Accurate insertion was confirmed with a 10 mL bolus of lidocaine via the catheter followed by bupivacaine (10 mL). If a continuous infusion was indicated, a portable pump was used to infuse ropivacaine (6–8 mL/h, 500 mL reservoir). Electric current was delivered using the integrated pulse generator for 7 days, followed by removal at home. Participants were contacted for data collection on postoperative days 1–5, 7, 8, and 14.ResultsDuring the first seven postoperative days, the median (IQR) “average” daily pain intensity as measured with the numeric rating scale (NRS) was 2 [0, 3]. Following postoperative day 1, daily average and maximum/worst NRS was less than 2 and 5, respectively. Regarding the “worst” (maximum) pain level over the entire study period of Days 0–14, 7 (35%) experienced mild pain (NRS<4), 10 (50%) moderate pain (NRS 4–7), and 3 (15%) severe pain (NRS>7). Cumulative oxycodone consumption the first week was a median of 43 (18, 73) mg.ConclusionsThis novel device successfully delivered a single-injection local anesthetic bolus, a continuous perineural infusion, and concurrent/subsequent electrical current. A randomized, sham-controlled clinical trial appears warranted.Clinicaltrials.gov NCT06818708 Introduction Postoperative pain is often undertreated due to lack of effective analgesics. Opioids have significant well-known limitations, while single-injection unencapsulated local anesthetic nerve blocks provide less than 24 hours of relief. Liposomal bupivacaine or continuous peripheral nerve blocks may extend analgesia to 48–96 hours.1 However, many surgeries result in pain lasting weeks rather than days. Percutaneous peripheral nerve stimulation (PNS) has emerged as a promising modality for extending postoperative analgesia while simultaneously reducing the need for opioids, without systemic adverse effects.2 However, the only PNS system currently cleared by the US Food and Drug Administration (FDA) for acute pain management (SPRINT PNS System, SPR Therapeutics, Cleveland, OH) was originally designed for chronic pain indications and costs several thousand dollars.3 As a result, despite ongoing investigation into its role in acute pain management, widespread clinical adoption remains unlikely. Furthermore, the analgesia provided by this system is insufficient to meet the demands of the immediate postoperative period, necessitating the addition of a single-injection local anesthetic peripheral nerve block, thus requiring two distinct procedures and separate sets of equipment.3 A potential solution currently under investigation for postoperative analgesia is the RELAY system (Gate Science, Moultonborough, NH), a novel device designed to integrate both local anesthetic delivery and PNS. The system uses a relatively standard catheter-over-needle design that permits ultrasound-guided percutaneous placement, with the catheter remaining in the perineural space following the withdrawal of the internal needle (figure 1). An initial single-injection bolus of local anesthetic can be administered rapidly at the time of insertion via either the needle or catheter. Postoperatively, the catheter may be connected to a portable infusion pump to deliver a continuous perineural local anesthetic infusion. Distinct from the existing perineural catheter systems, the RELAY catheter incorporates an embedded wire and integrated pulse generator with a battery capable of stimulating the targeted peripheral nerve(s) for up to 28 days. A mobile application facilitates a Bluetooth connection with the pulse generator, thereby enabling patients to adjust the stimulation intensity using a personal smartphone or tablet. The RELAY system (Gate Science, Moultonborough, New Hampshire) is a traditional catheter-over-needle device to allow the administration of a single injection of local anesthetic via the needle (or catheter) followed by a perineural local anesthetic infusion via the remaining catheter (when desired). The device is available with a short (7.7 cm) or long (11.5 cm) needle and 7.3 cm or 11.4 cm catheter, respectively. The catheter incorporates an embedded wire and integrated pulse generator capable of stimulating the targeted peripheral nerve(s) for up to 28 days. Each catheter has three electrodes at the end, which may be programmed as the anode or cathode. Small projections are set back approximately 4 cm from the catheter tip; and when the catheter is rotated counterclockwise, these projections are at approximately 70°–90° to the central axis of the catheter, while clockwise rotation brings these projections flush with the catheter for removal. This device has the potential to provide sustained analgesia for several weeks following a single percutaneous procedure. Such an approach may represent a paradigm shift in postoperative pain management by enabling the delivery of (1) a single-injection peripheral nerve block, (2) a continuous peripheral nerve block, and (3) neuromodulation—all through a single system that can be placed in a timeframe comparable with that of a traditional single-injection block. Rather than limiting analgesia to less than 24 hours, this integrated approach could extend pain control for up to 28 days, without requiring significant changes to current clinical workflows. We, therefore, conducted a prospective feasibility study under an FDA Investigational Device Exemption to assess the possibility of using the device to treat postoperative pain and optimize both the insertion technique and pulse generator parameters. Methods This study followed good clinical practice and was conducted within the ethical guidelines outlined in the Declaration of Helsinki. The device under study is not cleared for general use by the US FDA and, therefore, the protocol and an Investigational Device Exemption (G230193) were approved by the FDA prior to enrolling the first participant. The protocol was also approved by the local Institutional Review Board (University of California San Diego, La Jolla, California). Written, informed consent was obtained from all participants. The study was prospectively registered at clinicaltrials.gov (NCT06818708, PI: Brian Ilfeld, first registered 7 February 2025). The first participant was enrolled on 20 March 2025. Participants Enrollment was offered to adult patients at least 18 years of age scheduled for primary, unilateral orthopedic surgery of the shoulder, ankle, or foot anticipated to have moderate-to-severe pain for at least 1 postoperative week with a scheduled single-injection peripheral nerve block(s). Typical procedures included rotator cuff repair, total shoulder arthroplasty or ankle arthroplasty/arthrodesis. While surgeries not involving the shoulder, foot or ankle may induce significant pain, we chose these procedures because it has been demonstrated that percutaneous PNS can successfully deliver analgesia in these cases.2 An available smartphone able to download the pulse generator control application was also required. Patients were excluded for(1) chronic opioid use inclusive of tramadol (daily use within the 2 weeks prior to surgery and duration of use>4 weeks); (2) neuromuscular deficit of the surgical limb; (3) history of substance misuse; (4) implanted electric stimulator (eg, cardiac pacemaker); (5) history of bleeding disorder; (6) antiplatelet or anticoagulation therapy; (7) allergy to skin-contact materials (eg, occlusive dressings); (8) immunocompromised state; (9) incarceration; (10) pregnancy; (11) inability to communicate with or contact the investigators during the treatment period, and vice-versa; (12) allergy to amide local anesthetics; or (13) morbid obesity (body mass index>40 kg/m2). Intervention The pulse generator was turned on by pressing a single button on the device and connected via Bluetooth to a mobile device (iPhone 16, Apple Inc., Cupertino, CA) using a proprietary application (Gate Keeper, Gate Science, Moultonborough, NH). The minimum and maximum currents were set to 0.2 mA and 10.0 mA, respectively. The catheter has three separate electrodes proximal to the orifice with the pulse generator initially programmed to use the most distal as the anode and the middle as the cathode (the most proximal electrode was inactive). Preoperatively, participants had standard external monitors placed and oxygen delivered by facemask. The peripheral nerve block site(s) were cleansed with chlorhexidine gluconate and isopropyl alcohol and covered with a sterile fenestrated drape. Intravenous sedation/analgesia with midazolam and fentanyl was titrated for patient comfort while ensuring that participants remained responsive to verbal cues. The nerves treated included the sciatic proximal to the bifurcation for foot/ankle surgery and the brachial plexus for shoulder surgery. The target nerve(s) were visualized by ultrasound with the transducer placed within a sterile sleeve. For lower extremity surgery, a cross-sectional view of the sciatic nerve was imaged. For upper extremity surgery, two approaches were used. For the first, a transverse cross-sectional (short axis) view of the brachial plexus roots was obtained and the transducer rotated approximately 90°, resulting in an oblique view of the C5/C6 nerve roots with the transducer-oriented roughly cephalad–caudad. Alternatively, a more-traditional posterior approach to the brachial plexus trunks was used.4 While ensuring that the entry point was posterior enough to avoid interfering with the surgical drapes or directly under any postoperative arm sling,5 a skin wheal of lidocaine was raised immediately superior to the transducer to anesthetize the skin and then the track toward the target nerve(s). For lower extremity surgery, the device was advanced from the lateral aspect of the thigh to a subepimyseal extraparaneural location posterior to the sciatic nerve with small boluses of D5W (5 mL) via the integrated needle to open the potential space. For upper extremity surgery, the device was advanced either between the superior and middle trunks or immediately posterior to the C5/C6 nerve roots (online supplemental figure) with small boluses of D5W (5 mL) via the integrated needle to open the potential space between the roots and surrounding fascia. The pulse generator was initiated at 0.2 mA (100 Hz, 100 μs, biphasic rectangular waveform) and the current increased until the patient perceived sensory changes—usually described as “tingling”—in the surgical area. The 316 stainless steel internal guide needle was removed and lidocaine 2% with epinephrine (10 mL) injected via the remaining catheter. Accurate catheter tip placement was demonstrated with a decreased cutaneous sensation to cold temperature over the deltoid muscle or on the plantar aspect of the foot for shoulder or foot/ankle surgeries, respectively. After confirming a successful block, bupivacaine 0.5% with epinephrine (10 mL) was subsequently injected to maximize the single-injection peripheral nerve block duration. The catheter was rotated counterclockwise to help anchor the device with small projections (“tines”) emanating from the catheter at approximately 4 cm from the distal end (figure 1). These small projections are set back approximately 4 cm from the catheter tip; and when the catheter is rotated counterclockwise, these projections are at approximately 70°–90° to the central axis of the catheter, while clockwise rotation brings these projections flush with the catheter for removal. The device was secured with wound closure adhesive (2-Octyl 2-cyanoacrylate) and benzoin, a chlorhexidine-impregnated biopatch applied around the catheter entry point, and an overlying occlusive dressing (figure 2). The RELAY system (Gate Science, Moultonborough, New Hampshire) is seen following insertion adjacent to the nerve roots of C5 and C6 for shoulder surgery. For participants who were scheduled to receive a continuous peripheral nerve block (technique selection individualized by the anesthesia team in concert with the surgical team), ropivacaine 0.2% infusion was initiated with a portable infusion pump (ambIT, Avanos Medical, Alpharetta, GA) via the catheter as soon as feasible (8 mL/h basal infusion, 4 mL bolus dose, 30 min lockout period, and 500 mL reservoir). For ankle surgery, a single-injection bupivacaine 0.5% with epinephrine (10 mL) adductor canal block was administered using ultrasound guidance and a standard Tuohy needle. Participants had a general anesthetic of inhaled anesthetic combined with nitrous oxide and opioids administered at the discretion of the intraoperative anesthesia team for their surgical procedure. Within the recovery room, the patient’s phone was connected to the pulse generator via a Bluetooth connection. The proprietary controlling application allowed adjustment of the amplitude between 0.2 and 10 mA. Patients were discharged home with verbal and written instructions as well as a prescription for oxycodone (5 mg tablets), additional occlusive dressings to redress the device, extra gauze and sterile tubing caps for participants with a continuous peripheral nerve block. Patients were instructed to take acetaminophen 975–1000 mg four times per day, celecoxib 200 mg two times per day, and, if needed, oxycodone (5 mg tablets). Participants were permitted to shower with the pulse generators active and to subsequently change the dressing, if desired. Participants with a continuous peripheral nerve block could choose to return to the center to have their tubing disconnected from the neuromodulation device by an investigator following local anesthetic reservoir exhaustion or perform this procedure at home with instructions provided over the telephone by an investigator. On postoperative day 7, participants or their caretakers removed the occlusive dressing and withdrew the catheter (and integrated electrodes) with gentle traction and clockwise rotation—which reversed the catheter’s anchoring projections—at home with instructions provided over the telephone by an investigator. The devices are single use and disposable. Following study completion, the results were provided to all participants using non-technical language. Outcome measurements (endpoints) Participants were contacted by telephone for endpoint collection on postoperative days 0, 1, 2, 3, 7, 8, and 14. We selected outcome measures that have established reliability and validity, with minimal inter-rater discordance, and are recommended for pain-related clinical trials by the WHO and the Initiative on Methods, Measurement, and Pain Assessment in Clinical Trials (IMMPACT) consensus statement.6 Although there was no comparison group for this feasibility study, in preparation for a subsequent randomized, sham-controlled trial, we designated as the primary outcome measure the median “average” daily pain intensity for the first postoperative week (Days 1–4 and 7) as measured with the numeric rating scale (NRS). The NRS is a highly sensitive measure of pain intensity with numbers ranging from 0 to 10, zero equivalent to no pain and 10 equivalent to the worst imaginable pain; it is a valid and reliable measure for evaluating analgesic interventions.7 Additionally, NRS scores correlate well with other measures of pain intensity8 and demonstrate high test–retest reliability.9 These NRS characteristics led to WHO and the IMMPACT consensus recommendations for use of the ten-point NRS of pain intensity for pain trials.6 Additional outcomes included daily opioid consumption, average pain, worst pain and awakenings due to pain. The brief pain inventory (BPI, short form) assesses pain and its interference with physical and emotional functioning and was administered on postoperative days 3 and 7.10 The instrument includes three domains: (1) pain, with four questions using an NRS to evaluate four pain levels: “current,” “least,” “worst,” and “average”; (2) percentage of relief provided by pain treatments with one question (not used for this study); and, (3) interference with physical and emotional functioning using a 0–10 scale (0=no interference and 10=complete interference). The seven interference questions involve general activity, mood, walking ability, normal work activities (both inside and outside of the home), relationships, sleep, and enjoyment of life.10 These seven functioning questions can be combined to produce an interference subscale (0–70). The use of both single items (eg, mood) and the composite scores is supported by the IMMPACT consensus recommendations for assessing pain in clinical trials.6 11 Statistical analysis This investigation was designated a priori as a pilot feasibility study to assist in planning a subsequent randomized trial and we, therefore, used a convenience sample of 20 participants. Statistics were not applied to the data. Results Between 20 March 2025 and 12 June 2025, a total of 20 participants were enrolled (table 1) and had a neuromodulation device inserted (figures2 3) with sensory changes in the surgical area with activation of the pulse generator (tables1 2). Local anesthetic bolus with lidocaine 2% (10 mL) resulted in decreased sensation to cold temperature in 19 of 20 participants (95%) with the first attempt at catheter insertion. In the remaining participant (5%), stimulation was sensed in the shoulder region when current was initially delivered via the lead inserted adjacent to the brachial plexus (cephalad–caudad approach), but the catheter was apparently dislodged during needle removal as the subsequent local anesthetic was viewed collecting superficial to the plexus during injection under ultrasound. The catheter was removed, and a new device was inserted with a more-traditional posterior–anterior approach with successful injection of local anesthetic. Of the 20 participants who experienced sensory changes in the anticipated nerve distributions within 30 min, 7 (35%) had both a complete sensory and motor block, while the remainder had a partial sensory and motor block when the bupivacaine bolus was administered, after which no further testing of the block was performed. Table 1Baseline and procedural characteristicsCharacteristicAll(n=20)Single injection(n=8)Continuousinfusion(n=12)Age (years)57 (14)51 (12)60 (14)Female, n (%)9 (45%)2 (25%)7 (58%)Height (cm)173 (9)176 (10)171 (8)Weight (kg)87 (20)93 (26)83 (14)Body mass index (kg/m2)30 (6)30 (7)28 (5)Lead location, n (%) Brachial plexus15 (75%)7 (87%)8 (67%) Sciatic5 (25%)1 (13%)4 (33%)Surgical laterality (left), n (%)4 (20%)2 (25%)2 (17%) Primary surgical procedure Total shoulder arthroplasty7 (35%)0 (0%)7 (35%) Rotator cuff repair7 (35%)6 (75%)1 (8%) Clavicle open reduction internal fixation1 (5%)1 (13%)0 (0%) Ankle arthrodesis or arthroplasty1 (5%)0 (0%)1 (8%) Other ankle procedure4 (20%)1 (13%)3 (25%)Surgery duration (min)76 (40)63 (45)85 (35)Tourniquet duration, in four participants all with a continuous peripheral nerve block (min)91 (41)N/A91 (41)Continuous peripheral nerve block discontinuation (postoperative day)2.5 (1.0)N/A2.5 (1.0)Day of discharge, n (%) Postoperative day 017 (85%)7 (87%)10 (83%) Postoperative day 13 (15%)1 (13%)2 (17%)Values are reported as mean (SD) or number of subjects (percentage).N/A, not applicable. Baseline and procedural characteristics Values are reported as mean (SD) or number of subjects (percentage). N/A, not applicable. Table 2Electrical stimulation and injectionCharacteristicAll(n=20)Single injection(n=8)Continuousinfusion(n=12)Minimum current sensed (mA)2.5 (1.3)2.0 (0.4)2.8 (1.6)Maximum current comfortable (mA)3.8 (2.0)3.3 (0.6)4.2 (2.5)Electrode configuration, n (%) Anode (distal)20 (100%)8 (100%)12 (100%) Cathode (middle)6 (30%)1 (13%)5 (42%) Cathode (proximal)14 (70%)7 (87%)7 (58%)Muscle contraction present, n (%)6 (30%)2 (25%)4 (33%)D5W injection volume (mL)4 (5)2 (1)5 (6)Time from needle insertion to local anesthetic injection (min)9 (7)8 (4)9 (9)Time from needle first in to last out (min)10 (4)11 (4)9 (4)Fentanyl (µg)70 (37)78 (31)65 (41)Midazolam (mg)1 (1)1 (1)1 (1)Intraprocedural pain (NRS)Highest pain score (NRS)2.7 (1.8)3.0 (1.8)2.4 (1.9)Average pain score (NRS)1.1 (1.2)1.0 (0.8)1.1 (1.4)Values are reported as mean (SD) or number of subjects (percentage).NRS, numeric rating scale. Electrical stimulation and injection Values are reported as mean (SD) or number of subjects (percentage). NRS, numeric rating scale. Effects of a single-injection peripheral nerve block, a continuous ropivacaine 0.2% perineural infusion (in select participants), and 7 days of percutaneous PNS on daily average and worst/maximum pain intensity as well as opioid consumption for all procedures combined. Pain intensity was measured using an NRS with 0 equal to no pain and 10 being the worst imaginable pain. Opioid consumption is measured in oxycodone equivalents. Left-side panels include all participants, while the right-side panels present participants with (1) sciatic lead; (2) brachial plexus lead with a continuous peripheral nerve block; and (3) brachial plexus lead without a continuous peripheral nerve block (cPNB). Data expressed as median (dark horizontal bars) with 25th–75th (box), 10th–90th (whiskers), mean (diamonds), and outliers (circles). NRS, numeric rating scale; PNS, peripheral nerve stimulation. During the first seven postoperative days, the median (IQR) “average” daily pain intensity as measured with the NRS was 2 (0, 3). Following postoperative day 1, daily average and maximum/worst NRS had a decreasing trajectory and were equal to or less than 2 and 5, respectively (figure 3). Regarding the highest “worst” (maximum) pain level over the entire study period of Days 0–14, 7 (35%) experienced mild pain (NRS<4), 10 (50%) moderate pain (NRS 4–7), and 3 (15%) severe pain (NRS>7). Cumulative oxycodone consumption the first week was a median of 43 (18, 73) mg (figure 3). The BPI measuring physical and emotional interference due to pain was 2 (0, 4) on Day 3 and 0 (0, 7) on Day 7. Cumulative awakenings due to pain were a median 2 (0, 7). Participants with a postoperative continuous peripheral nerve block reported sensory changes in the surgical limb each day during the local anesthetic infusion, and none more than 4 hours following reservoir exhaustion. Adverse events and protocol deviations No serious device-related adverse events or systemic side effects were identified. For one participant, during baton rotation prior to intentional withdrawal, the pulse generator separated from the catheter, but the catheter was easily manually extracted by the patient’s caretaker. Infusion pump occlusion alarms sounded prior to discharge in three participants, and the blockage could be partially cleared in only one with the remaining two pump tubing sets discontinued. The remaining participant’s pump alarmed at home after infusing 220 mL and the tubing set was discontinued prematurely. One participant had his catheter/lead accidentally dislodged during the removal of the continuous peripheral nerve block tubing from the lead on postoperative day 3. Another participant removed the catheter/lead at the end of the perineural infusion on postoperative day 2 because she “did not want to deal with it.” Finally, one participant disclosed a history of substance misuse—an exclusion criterion—after reporting a pain score of 10 in the recovery room, which resolved with a bolus of local anesthetic, and we, therefore, chose to send him home with a continuous infusion, which had not originally been planned for this patient. Discussion This pilot study suggests that a novel percutaneous device successfully integrates both local anesthetic delivery and PNS, and its outpatient use following surgery is feasible. Such an approach may represent a paradigm shift in postoperative pain management by enabling the delivery of (1) a single-injection peripheral nerve block, (2) a continuous peripheral nerve block, and (3) electrical neuromodulation—all through a single system that can be placed in a timeframe comparable with that of a traditional single-injection nerve block. The only block failure was experienced during our third brachial plexus insertion in which we were using a cephalad-to-caudad approach; and this patient ultimately had a successful local anesthetic block using a posterior-to-anterior approach. After reverting to this more-traditional technique for subsequent insertions, no further block failures were observed. The first participant to receive a continuous peripheral nerve block had a catheter found to be occluded in the recovery room and which could not be cleared. In our second such participant, we initiated the infusion immediately following device insertion finding a patent catheter; but the infusion was held in the operating room and was subsequently found to be occluded in the recovery room. It could only be partially cleared with a hand-delivered bolus (and occluded at home after infusion of 225 mL). A third participant had an infusion begun following insertion but had an occlusion that could not be cleared within the recovery room. We theorize that the three exposed electrodes at the distal end of the catheter catalyzed hemostasis resulting in an occluded catheter. We found that setting the electrical current at a relatively low setting (1–2 mA) and instructing participants to increase the current when their postoperative pain increased during local anesthetic block resolution helped to avoid an uncomfortably high current. While the time of insertion was relatively high for the initial five participants, this time decreased precipitously after familiarization with the equipment so that the catheters for our final five participants were placed in a timeframe comparable with that of a traditional single-injection nerve block. The participant with undisclosed substance misuse who experienced severe pain in the recovery room highlights an unexpected benefit of the current device; while we had originally planned on discharging this patient solely using neuromodulation to treat pain following resolution of the single-injection peripheral nerve block, we were able to convert this to a continuous peripheral nerve block due to the integrated catheter–lead system. The most important weaknesses of this feasibility study were the limited sample size and lack of a control group. A randomized, sham-controlled study will be required to adequately evaluate the analgesia potential of this novel device and appears warranted. Furthermore, the success of the device in inducing surgical anesthesia remains unknown because we injected the bupivacaine bolus via the catheter as soon as any sensory deficit was detected following the initial lidocaine bolus. While only seven (35%) of participants experienced a complete sensory and motor block when we first detected the sensory block, all or a subset of the remaining participants could have developed complete blocks without identification. Finally, as with most feasibility studies, the exclusion criteria resulted in a rather homogeneous sample of relatively healthy individuals to optimize internal validity but decrease external validity; our results cannot be generalized to subpopulations with chronic pain, morbid obesity, and other criteria for which we excluded. Supplementary material