Compression therapy using normal-sized surgical gloves for prevention of paclitaxel-induced peripheral neuropathy in breast cancer: Results from the multicenter randomized AIUR trial

AIUR

Patient / Population Intervention / Exposition Comparaison Critère de jugement

In patients with stage II-III breast cancer receiving paclitaxel, compression with normal-sized surgical gloves did not reduce patient-reported worsening of peripheral neuropathy compared with no intervention.

N

104patients

Schéma

Multicentre, parallel-group trial at seven centres in South Korea randomising patients 1:1 with stage II-III breast cancer receiving at least 12 weeks of paclitaxel to two normal-sized surgical gloves worn from 30 min before to 30 min after each infusion or to no intervention (NCT05771974).

Critère

Proportion of patients with a FACT-NTX 11-item increase of at least 5 points at the end of paclitaxel.

RésultatThe primary endpoint was not met: poor outcome 74.5% with gloves vs 82.6% with no intervention, difference -8.1 (95% CI -24.7 to 8.5; P = 0.45). Change in FACT-NTX total score +9.7 vs +13.7, difference -3.9 (95% CI -7.5 to -0.4; P = 0.03). Upper-extremity sub-score change +2.9 vs +6.1, difference -3.20 (95% CI -4.93 to -1.48; P < 0.001). Lower-extremity sub-score change +5.3 vs +5.5, difference -0.22 (95% CI -1.69 to 1.25; P = 0.77). Arm-by-region interaction P < 0.001. CTCAE grade ≥2 sensory neuropathy 58.7% vs 82.2% (P = 0.02). No serious adverse events. 44 of 48 assessable patients (92%) rated the gloves satisfactory or better.

Kang Y, Yoon CI, Yang Y, et al. Presented at ESMO Congress 2026, abstract LBA28, 23 October. Not yet published.

Discussion et critique

AIUR did not meet its patient-reported primary endpoint, so it is a negative trial. The signal in the gloved hands, together with no effect in the ungloved feet, is biologically coherent, but it rests on exploratory regional sub-scores and secondary endpoints that were not prespecified as primary. The abstract does not describe blinding. With a no-intervention control, patients knew their allocation, which can bias both patient-reported and investigator-rated neuropathy scores. With 104 patients the trial is small, the confidence interval for the primary endpoint is wide, and outcomes are reported only at the end of paclitaxel, with no data on longer-term persistence of neuropathy.