Transcutaneous Electrical Nerve Stimulation as an Adjunct to Local Anaesthesia During Transperineal MRI-Ultrasound Fusion-Guided Prostate Biopsy: A Randomized Triple-Blind Sham-Controlled Trial

This is a single-centre, prospective, randomized, sham-controlled, two-parallel-group superiority trial. It builds on a preceding single-centre randomized sham-controlled pilot (TENS 1, n=84, 1:1:1 allocation to active TENS, sham TENS and local anaesthesia alone), which demonstrated large effect sizes for TENS-associated pain reduction during periprostatic nerve block and biopsy sampling, with no difference between sham and local anaesthesia alone at any stage and no device-related adverse events. The pilot was retrospectively registered, was explicitly hypothesis-generating and had no formal power calculation. The present trial addresses each of these limitations: it is prospectively registered, formally powered against a pre-specified minimum clinically important difference, and restricted to two arms on the basis of the pilot finding that sham stimulation is an inert control condition.

Stimulation is delivered through a single channel using two electrodes placed paraperineally, one on each side, so that the current path crosses the perineum transversely over the S2 to S4 dermatomes, corresponding to the pudendal innervation of the perineum and the periprostatic region. The single-centre design is deliberate: all biopsies are performed at one institution using the same fusion platform, the same local anaesthesia technique and the same stimulation protocol as the pilot, which minimises inter-operator and inter-centre variability and preserves methodological continuity, at an acknowledged cost to external validity.

Pain scores are self-recorded by the participant on a paper form at fixed procedural stages, each announced in standardised wording by the operator, who is blinded to allocation. No member of the study team asks the participant for a pain score at any point. This differs from the pilot, in which scores were collected verbally, and the change is accounted for when pilot and trial effect sizes are compared.

Sample size is not based on the pilot effect estimate, because effect sizes from small pilot trials are systematically inflated. The planning scenario uses the pre-specified minimum clinically important difference of 1.5 NRS points with an assumed standard deviation of 2.0 (Cohen's d = 0.75), a two-sided alpha of 0.05, a Mann-Whitney U test and an asymptotic relative efficiency adjustment of 0.955. The trial adopts 60 evaluable participants per group, which provides 80% power for effects of d = 0.52 or larger and retains 76% power in a conservative robustness scenario of d = 0.50. Recruitment is inflated for dropout and non-evaluable primary outcomes. No interim analysis is planned.

The primary analysis is by intention-to-treat and uses a two-sided Mann-Whitney U test, with the Hodges-Lehmann estimate of median difference and 95% confidence interval reported alongside rank-biserial correlation and epsilon-squared for comparability with the pilot. A per-protocol sensitivity analysis excludes major deviations. A primary outcome missing because of discontinuation before the nerve block is handled by worst-case imputation, with best-case imputation as a secondary sensitivity analysis. Secondary analyses are presented without multiplicity correction and interpreted as supportive rather than confirmatory. Blinding integrity is quantified using the James index reported overall and the Bang index reported separately for each arm; these analyses are descriptive and do not modify the primary analysis. A separate statistical analysis plan is locked before database lock and unblinding.