By Victor Pikov, contributing editor
September 8, 2026 | UK-based Galvani Bioelectronics is working to develop a rheumatoid arthritis therapy using splenic nerve stimulation (SpNS). Initial results from the first 11 patients were published in The Lancet this month (DOI: 10.1016/j.eclinm.2026.104133).
Galvani’s proposed therapy builds on the VNS approach for suppressing splenic overactivity, originally pioneered by SetPoint Medical. While SetPoint applies VNS at the main cervical vagal trunk, Galvani hypothesized that targeting the specific branch innervating the spleen would deliver greater efficacy with fewer off-target-effects, such as bradycardia. To test this hypothesis, Galvani first had to develop a novel neuromodulation device for the splenic nerve, which is situated approximately 10 cm deep and adjacent to the pancreas.
Galvani was spun out from GSK in 2016 following co-investment from Verily, an Alphabet company. (For a full overview of Galvani’s journey from inception to its recent shutdown preparations, see our April 2026 BBR article.) Galvani collaborated closely with a Verily team led by Peng Cong (now at Echo Neurotechnologies) to engineer a miniature IPG capable of wireless charging at a depth of 2.5 cm—compared to the typical 1–2 cm range—and transferred the design to Integer (soon to be acquired by KKR) for FDA-compliant manufacturing. Targeting the splenic nerve also required Sebastien Ouchouche at Galvani (now at INBRAIN Neuroelectronics) to develop an innovative cuff that wraps around the splenic neurovascular bundle, as individual splenic nerve fascicles are too small and deeply embedded within the arterial wall to target independently.
In 2021, with the IPG and cuff ready for testing, Galvani conducted a pig study to identify a real-time biomarker for tuning SpNS parameters, bypassing the significant delay required to observe anti-inflammatory endpoints. A team led by Matteo Donegà (now also at INBRAIN Neuroelectronics) demonstrated that reduced splenic arterial blood flow served as a rapid biomarker: it responded within 30 seconds, showed sensitivity to changes in SpNS amplitude and frequency, and correlated with pro-inflammatory cytokine levels during acute inflammation. Donegà’s team also confirmed that SpNS—despite the splenic nerve being sympathetic—elicited the same anti-inflammatory effect as VNS, which is notable given that sympathetic stimulation typically opposes vagal actions. A definitive mechanism for this vagal-like response remains unconfirmed; rodent anatomical studies suggest the vagus nerve may modulate splenic activity via synaptic connections in the celiac ganglion, though corresponding anatomical data in humans and pigs are lacking.
Building on these preclinical findings, Galvani and clinical collaborators at Catharina Hospital in the Netherlands published results from a 2022 first-in-human study. In that study, acute SpNS was delivered to 13 patients undergoing esophagectomies near the splenic nerve, confirming that SpNS reduced splenic blood flow in humans consistent with the pig model.
Following confirmation of acute safety and performance, Galvani launched a pilot chronic clinical study evaluating the combined cuff and IPG system. A total of 16 patients were enrolled across 14 sites in the U.S. and the Netherlands before Galvani began shutdown preparations in 2026. This month’s Lancet publication includes results from the first 11 patients across 6 sites (implanted between 2022 and 2024 and continuing follow-up). During screening, 32 of 44 candidates were excluded, primarily because high BMI and abdominal wall thickness prevented IPG placement within the 2.5 cm depth limit required for transcutaneous charging.
These pilot findings offer a useful comparison to SetPoint’s pivotal RESET-RA study published in January 2026 (Nature Medicine, DOI: 10.1038/s41591-025-04114-7). SetPoint enrolled 242 patients with moderate-to-severe RA who demonstrated intolerance or inadequate response to at least one disease-modifying antirheumatic drug (DMARD). In contrast, Galvani enrolled patients who had failed at least two DMARDs with distinct mechanisms of action (typically anti-TNF and JAK inhibitors). Both studies utilized a three-month sham-control phase in accordance with FDA Guidance which ethically limits withholding active treatment to three months. SetPoint’s study continued monitoring VNS efficacy beyond three months, showing that peak efficacy required six to 12 months to build up. Unfortunately, Galvani missed the opportunity to evaluate SpNS efficacy beyond three months because the protocol required adding pharmacotherapy after 12 weeks if disease activity remained moderate or high.
Clinical efficacy data from Galvani’s SpNS cohort remain difficult to interpret, as only three patients in the treatment arm and one patient in the sham arm completed the three-month randomized phase. When constrained by small sample sizes, pilot studies typically use subjects as their own historical controls, an approach Galvani also omitted. With these caveats, two primary efficacy metrics can be compared. On the DAS28-CRP score (a composite score of RA symptoms, serum C-reactive protein, and global health assessments), both Galvani’s SpNS cohort and SetPoint’s VNS cohort achieved a mean reduction of 1.1, compared to sham reductions of 0.3 and 0.7, respectively. On the ACR20 endpoint (the proportion of patients achieving at least 20% symptom improvement), Galvani’s SpNS arm achieved a 44% response rate, while SetPoint’s VNS arm reached 35%. In SetPoint’s sham arm, 24% of patients improved, whereas the single evaluable sham patient in Galvani’s study improved, resulting in a nominal 100% sham response rate.
Adverse events occurred in 27% of patients with Galvani’s abdominally implanted device (including implant-site pain, dyspnea, procedural nausea, defecation urgency, hunger, abdominal discomfort, and hyperhidrosis) with zero serious adverse events observed in this small cohort. In SetPoint’s larger pivotal study, 21% of patients receiving the cervical implant reported adverse events (including hoarseness, inflammation, numbness, suture infection, implant-site pain, and erythema), with 1.6% experiencing serious adverse events.
In Galvani’s study, most SpNS-treated patients experienced uncomfortable sensations, graded as mild (40%), moderate (21%), or severe (2%). These sensations constrained stimulation amplitude, requiring multiple clinical visits to titrate patients to a maximum tolerable range of 3-9 mA, substantially below the 8-20 mA optimal window identified during acute human testing. In contrast, SetPoint’s VNS therapy was well tolerated, with stimulation amplitudes routinely titrated up to a maximum of 2.5 mA over time.
Implanting the Galvani system required a complex laparoscopic procedure for both the IPG and cuff, averaging 139 minutes across 11 patients. Conversely, SetPoint’s integrated cuff-and-IPG device required a simpler open surgical approach at the cervical vagus nerve lasting 60-90 minutes.
Overall, both neuromodulation platforms demonstrated acceptable safety profiles. However, Galvani’s core hypothesis—that target-specific stimulation would yield superior efficacy—remains unsupported by these limited results. Galvani’s approach was primarily constrained by off-target abdominal sensations, strict anatomical charging limits, and a longer surgical procedure. Furthermore, as noted in our April 2026 BBR article, an elevated corporate valuation hindered Galvani from securing the external capital required to complete this pilot and fund a pivotal PMA study. As former team members transition across the bioelectronic medicine industry, the insights and technical knowledge gained throughout the Galvani journey will help inform future neuromodulation endeavors.
