The correspondence arrives at a moment when interventional pain procedures are being scrutinized with unprecedented rigor. Pulsed radiofrequency, the technique evaluated in the original trial, applies short bursts of electrical current to nervous tissue, generating heat intermittently rather than continuously. Unlike conventional thermal radiofrequency lesioning, which deliberately destroys nerve fibers at sustained high temperatures, pulsed radiofrequency was developed on the premise that it can modulate pain signaling while sparing the treated structures from outright injury. That theoretical advantage has made it attractive for patients whose life expectancy and quality of life make destructive neurolytic procedures less appealing, and spinal metastases represent precisely such a population: people whose pain is often refractory to escalating opioid doses but whose remaining time is better spent awake and functional than sedated.

The problem that Chunbo Li and Ding Li highlight is not the biology of the technique but the epistemology of the evidence. In a randomized controlled trial comparing oxycodone plus pulsed radiofrequency against oxycodone alone, patients in the intervention arm know they received a procedure. They experienced the preparation, the local anesthesia, the positioning, and the post-procedural observation. Patients in the control arm know they did not. This asymmetry may sound trivial, but decades of clinical research have demonstrated that expectations, conditioning, and the therapeutic ritual surrounding an intervention can themselves produce measurable analgesia. The placebo response in pain is among the most robust in all of medicine, engaging endogenous opioid systems, descending inhibitory pathways, and prefrontal cortical networks that can be recruited simply by the belief that something active has been done.

A sham control is the methodological answer to this problem. In a properly sham-controlled device trial, participants in the comparison group undergo a procedure that is indistinguishable from the real one to the patient—identical skin preparation, local anesthesia, and sensory experience—without delivering the therapeutic element of the intervention. Only when both arms share this experience can the specific effect of the device be separated from the nonspecific effects of receiving any procedure at all. The correspondence authors cite a 2023 review by Johnson and Goebel that examined exactly this challenge, and its title says a great deal: sham controls in device trials for chronic pain are, as the authors put it, tricky in practice. The difficulty is real. Ethicists and investigators must weigh whether exposing control patients to needle insertion, sedation, or even brief electrical stimulation is justified by the knowledge gained, and institutional review boards often balk at sham procedures that carry any procedural risk.

This is why many published trials of interventional pain techniques, including some widely cited ones, compare active treatment against standard care or against another active procedure rather than against a credible sham. Such designs are easier to approve and easier to recruit for, but they answer a different question. They tell us whether the new treatment is better than what came before, not whether the treatment works through its proposed mechanism. When the comparison is unblinded standard care, the measured difference bundles together the true biological effect of the intervention, the placebo effect of undergoing a procedure, the nocebo effect of not undergoing one, and performance bias from clinicians who know which patients received the active treatment and may unconsciously titrate medications or assess pain differently.

The stakes are amplified in cancer pain research because the outcomes are almost always patient-reported. The original trial used the Numerical Rating Scale, a standard instrument in which patients rate their pain intensity from zero to ten. Patient-reported outcomes are the right primary endpoints for pain—there is no blood test for suffering—but they are also the endpoints most vulnerable to expectancy effects. Efficace and colleagues, in a 2022 analysis published in the Journal of the National Cancer Institute, examined the impact of blinding on patient-reported outcome differences between treatment arms in cancer randomized controlled trials and found that the integrity of blinding materially shapes the observed differences. When patients can guess their assignment, reported outcomes diverge in ways that may not reflect pharmacology or device mechanics. In pain trials, where the outcome is a subjective number, an unblinded design is not a minor limitation; it is a structural threat to the validity of the conclusion.

Patients with spinal metastases may be especially susceptible to strong expectancy effects, and for reasons that deserve compassionate acknowledgment. These are people in severe, often escalating pain who have frequently exhausted simpler options. Being offered a procedure carries a message of hope and action, and hope in this context is not a confound to be dismissed but a genuine therapeutic force. The correspondence authors are not suggesting that hope is bad; they are suggesting that a trial design that cannot distinguish the analgesic contribution of hope from the analgesic contribution of pulsed radiofrequency cannot tell clinicians what to expect when the technique is deployed in routine practice, where every patient will know they are being treated.

There is also a practical dimension to the critique. Refractory cancer pain from spinal metastases is a devastating condition, and clinicians need to know whether adding an interventional procedure to optimized opioid therapy delivers benefit that justifies its cost, its risks, and the burden of referral to a pain intervention service. If the observed benefit in an unblinded trial is inflated by expectancy, patients may be subjected to procedures with real complication profiles—infection, bleeding, nerve injury, and the general risks of any invasive intervention—for gains that would not survive a blinded comparison. Conversely, if pulsed radiofrequency genuinely works, an underpowered or biased literature may delay its adoption and leave patients on higher opioid doses than necessary, with the constipation, sedation, and immunomodulatory effects that chronic high-dose opioids entail. Good trial design is not academic pedantry; it is the mechanism by which patients receive treatments that actually help them.

The correspondence also touches on a broader tension in the device-trial literature. Surgical and procedural specialties have historically lagged behind pharmacology in adopting blinded designs, partly because blinding is genuinely harder to engineer. A drug trial can give every patient an identical-looking pill; a radiofrequency trial cannot give every patient an identical-looking lesion. Yet the past two decades have produced creative solutions: sham injections with saline, sham nerve blocks with inactive local anesthetic concentrations, sham stimulation devices that produce tingling without therapeutic current, and even sham surgery designs in which patients undergo arthroscopy or incision without the definitive repair. Each of these has been ethically contested, and each has reshaped the evidence base in its field. The lesson from those debates is that the question is not whether sham controls are perfect—they are not—but whether the alternative, an unblinded comparison, is good enough to support the clinical claims being made.

For readers of the original trial, the correspondence should prompt a careful rereading rather than a dismissal. The original study’s randomized design, its focus on a population with few good options, and its combination of a systemic opioid with a targeted procedure all represent genuine contributions to a sparse literature. Randomization, even without blinding, protects against selection bias and ensures that the groups being compared start from comparable baselines. The question raised by Li and Li concerns interpretation: how much of the reported between-group difference in pain scores reflects the specific neuromodulatory effect of pulsed radiofrequency, and how much reflects the psychological and contextual effects that accompany any procedure offered to a suffering patient.

The way forward, as the correspondence implicitly suggests, involves several converging efforts. Future trials in this space should, where ethics boards permit, incorporate sham procedures that mimic the sensory and procedural experience of pulsed radiofrequency without delivering it. Where sham designs are not feasible, investigators can strengthen blinding of outcome assessors, prespecify and measure patient expectations, use objective or semi-objective secondary endpoints such as opioid dose reduction, rescue medication use, sleep quality, and functional status, and report blinding-integrity checks asking participants to guess their assignment. Transparent reporting of these elements allows readers to calibrate their confidence in the results. Funders and journals, including those publishing supportive care and palliative research, increasingly expect such methodological detail, and correspondence like this one plays a constructive role in raising the bar.

It is also worth noting what the correspondence does not claim. The authors do not assert that pulsed radiofrequency is ineffective for spinal metastasis pain, nor do they dispute the clinical importance of the problem the original investigators tackled. Their intervention is methodological, and their acknowledgment of the original authors’ work is gracious, reflecting a scientific culture in which critique and gratitude coexist. The exchange is a model of how the literature should evolve: a trial reports a promising result, independent clinicians examine the design, and the field emerges with sharper questions and better-defined standards for the next study.

For patients and families navigating refractory cancer pain, the practical message is one of measured optimism. Multimodal care—combining optimized pharmacotherapy, interventional techniques where appropriate, radiotherapy for bone metastases, and comprehensive palliative support—remains the standard, and no single trial should upend that framework in either direction. For clinicians, the message is to weigh the evidence for procedural interventions with attention to whether the comparison group received a credible control. And for researchers, the message is that in pain medicine, where the outcome lives in the patient’s report and the treatment arrives wrapped in ritual and expectation, the sham control is not an optional refinement but the difference between measuring a treatment and measuring a hope. The correspondence by Li and Li is a concise, well-supported reminder of that distinction, published in a journal whose readers—oncologists, palliative care physicians, anesthesiologists, and nurses—stand to benefit from taking it seriously.

Subject of Research: Methodological critique of sham controls in a randomized trial of oxycodone plus pulsed radiofrequency for refractory cancer pain from spinal metastases

Article Title: Comment on “Oxycodone combined with pulsed radiofrequency for refractory cancer pain from spinal metastases: a randomized controlled study”

Article References: Li, C., & Li, D. (2026). Comment on “Oxycodone combined with pulsed radiofrequency for refractory cancer pain from spinal metastases: a randomized controlled study”. Supportive Care in Cancer, 34(11), Article 1122. https://doi.org/10.1007/s00520-026-11376-5

Image Credits: AI Generated

DOI: 10.1007/s00520-026-11376-5

Keywords: refractory cancer pain, pulsed radiofrequency, oxycodone, spinal metastases, sham control, randomized controlled trial, blinding, patient-reported outcomes, pain management, interventional pain therapy, clinical trial methodology, supportive care in cancer