Ear seeds ease chemotherapy nerve pain in trial, but only with in-person care

Chemotherapy can leave survivors with burning hands, numb feet, and tingling that persists for months or years, and clinicians have few tools to treat it. A randomized controlled trial now reports that auricular point acupressure, a technique in which tiny Vaccaria seeds are stuck to specific points on the outer ear and pressed several times a day, improved chemotherapy-induced peripheral neuropathy more than usual care. The advantage, however, was strikingly uneven: it appeared in both delivery formats during the first week, but only the arm that received in-person seed placement by a trained interventionist kept its edge over usual care beyond that point.

The trial, published in Supportive Care in Cancer, enrolled 227 cancer survivors who had completed neurotoxic chemotherapy at least three months before joining and who reported pain, numbness, or tingling in their extremities averaging at least 4 out of 10 during the previous week. Participants were randomized equally to one of three groups: in-person, mobile-supported acupressure with 80 participants; virtual, self-administered acupressure with 75; or usual care with 72. Both acupressure arms followed the identical four-week auricular point protocol and received one virtual coaching session a week after starting. The only difference was whether the initial seeds were placed in person or placed by participants themselves after standardized video instruction. Usual care meant standard oncologic and supportive treatment with no ear seeds and no acupressure training, though participants were told they could receive the intervention after the one-month control period.

Measuring neuropathy in a way that respects how differently it strikes each patient posed a design challenge. The researchers used the Modified Brief Pain Inventory–Chemotherapy-Induced Neuropathy scale, which rates pain, numbness, tingling, and stiffness across seven body regions from fingers to ankles on an 11-point scale. From the 28 possible symptom-by-location combinations, each participant’s single worst item at baseline was selected and then tracked over time, forming an individualized composite outcome anchored to the symptom that mattered most to that person. The primary endpoint was a clinically meaningful improvement, defined before analysis as a reduction of at least 20 percent from baseline in that individualized measure. Secondary outcomes included neuropathy interference with daily life, functional status, and quality of life, with the individual quality-of-life domains treated as exploratory.

The results, analyzed by intention to treat with generalized estimating equations and multiple imputation, showed a rapid but short-lived signal in symptom severity. At Week 1, both acupressure arms beat usual care: the in-person group had 2.60 times higher odds of achieving the 20 percent improvement threshold (95 percent confidence interval 1.27 to 5.33), and the virtual group had 3.22 times higher odds (95 percent confidence interval 1.48 to 7.01), both statistically significant. After that, the two arms diverged. The in-person advantage persisted at Week 2 (odds ratio 2.49, p = 0.02) and approached significance at Week 4 (odds ratio 2.05, p = 0.06), while the virtual arm’s advantage never again reached significance in any week. The authors note that the early improvement in both arms suggests nonspecific or behavioral factors contributed to the initial response, while the persistence of the in-person benefit points to an additional contribution from hands-on guidance.

Neuropathy interference, measured with the Brief Pain Inventory interference subscale and defined as a one-point reduction, told a more consistent story. Participants in the in-person arm had 3.97 times higher odds of meaningful improvement than controls (95 percent confidence interval 1.83 to 8.63), and the virtual arm had 3.15 times higher odds (95 percent confidence interval 1.47 to 6.76). Week-by-week comparisons showed the in-person benefit over usual care was significant at every week of the intervention, with odds ratios ranging from 3.29 to 4.28, all p < 0.01. The virtual arm’s interference benefit was significant at Weeks 2 and 4 but not at Week 3. No treatment-by-time interaction reached significance, meaning the effect neither grew nor shrank across the four weeks.

Exploratory quality-of-life findings added texture but also caution. In the in-person arm, anxiety, depression, and pain interference on the PROMIS-29 instrument improved significantly at multiple time points, with pain interference showing the largest effect at Week 3. In the virtual arm, these changes were smaller and largely non-significant. ECOG performance status did not differ between groups at any time point, and FACT/GOG-Neurotoxicity scores showed small, non-significant positive changes in both acupressure arms. No between-group differences emerged for fatigue or social functioning. The authors frame these domain-level results as exploratory rather than confirmatory.

The trial’s most serious weakness is differential attrition. Post-intervention completion was 74 percent in the in-person arm, 52 percent in the virtual arm, and 96 percent in usual care, an overall attrition rate of 26.4 percent. Non-completers were more often male, more often had high school education or less, and were disproportionately from the virtual arm; low educational attainment was the only independent predictor of dropping out. The authors argue the retention incentive offered to usual-care participants, who knew acupressure would follow their control period, likely explains their near-perfect completion, while the fully remote arm faced the same weekly assessment burden with no compensating human contact. Sensitivity analyses using delta-adjusted pattern-mixture models found the in-person arm’s severity advantage held up to a substantial departure from the missing-at-random assumption, but the virtual arm’s advantages tipped with only modest departures, meaning its conclusions are fragile.

The design also cannot isolate why the seeds might work. Because both acupressure arms received the same active point protocol, the comparison estimates the contribution of delivery modality, not of auricular point specificity. There was no sham condition and no attention control, participants and interventionists knew their assignments, and usual care involved no comparable contact. The authors state plainly that an equal-contact attention-controlled design is needed before the specific contribution of auricular stimulation can be established, and they regard this as the priority for the next trial. They also acknowledge that spontaneous recovery of neuropathy over time cannot be formally excluded as a contributor, though eligibility requirements placing participants at least three months past chemotherapy, and a mean time since diagnosis of about four years, make rapid natural resolution less likely to explain a four-week between-group difference.

Context matters for interpreting the size of these effects. Current American Society of Clinical Oncology guidance recommends no drug for preventing chemotherapy-induced peripheral neuropathy, and duloxetine remains the only agent with moderate evidence for treating established disease, producing a 0.6 to 1.0 point effect on a 10-point pain scale with no effect on numbness or tingling. Acupuncture delivered in person has shown benefit in a large pragmatic trial but requires repeated clinic visits, carries procedural risks such as tissue injury or infection, and is often not covered by insurance. Ear seeds, by contrast, cost a few cents per placement, need no needles or sterile technique, and require only a single training session plus one coaching contact. Adverse events in the trial were limited to anticipated local effects such as skin irritation and seed displacement, with no serious events attributed to the intervention in any arm.

The authors position auricular point acupressure as a candidate, not an established, component of multimodal survivorship care, and explicitly label these findings preliminary and hypothesis-generating rather than proof of efficacy. The severity advantage appeared at only one time point, attrition differed sharply across arms, and the trial lacked blinding and a sham comparator. What the data do support is a rapid-acting, low-cost, low-risk adjunct that may ease the daily-life interference of chemotherapy-induced peripheral neuropathy while it is being used, with the strongest and most durable signal coming when a trained interventionist places the seeds in person. Whether remote delivery can be redesigned with enough structured contact to match that benefit, and whether effects persist beyond the intervention period, remain open questions the authors direct to future work, including separately published follow-up analyses at one, two, and three months.

Subject of Research: Auricular point acupressure as a treatment for chemotherapy-induced peripheral neuropathy in cancer survivors

Article Title: Auricular point acupressure for chemotherapy-induced peripheral neuropathy: a randomized controlled trial

Article References: Lukkahatai, N., Jiang, Y., Zhang, J., Kawi, J., Wu, H., Wang, H., Campbell, C. M., Thrul, J., Sair, H., Hamilton, K. R., Cho, Y., Benjasirisan, C., Abdi, S., Hacker, E. D., Jafri, S. H., Cen, P., Gonzalez, A., Maithel, N., Rowe, J., … Smith, T. J. (2026). Auricular point acupressure for chemotherapy-induced peripheral neuropathy: a randomized controlled trial. Supportive Care in Cancer, 34(11), Article 1175. https://doi.org/10.1007/s00520-026-11373-8

Image Credits: AI Generated

DOI: 10.1007/s00520-026-11373-8

Keywords: chemotherapy-induced peripheral neuropathy, auricular point acupressure, ear seeds, randomized controlled trial, cancer survivors, neuropathy interference, virtual care delivery, Vaccaria seeds, supportive cancer care, PROMIS-29, differential attrition, nonpharmacologic symptom management