Extended-course transcutaneous tibial nerve stimulation for pediatric overactive bladder: a 6-Month prospective single-arm study

Abstract

Objective

To estimate the added clinical benefit of extending transcutaneous tibial nerve stimulation (TTNS) from 12 to 24 weeks in pediatric overactive bladder (OAB), and to characterize late responders.

Patients and Methods

Prospective single-arm cohort with home-based TTNS and assessments at baseline, 12, and 24 weeks. The primary outcome was change in OAB Symptom Score (OABSS) and responder status defined by a minimal clinically important difference (MCID) of≥3 points. Paired responder transitions were tested with McNemar’s exact test and summarized as paired risk difference (RD) with 10,000-sample bootstrap 95% CIs; repeated-measures GEE/LMM were pre-specified for confirmatory modeling.

Results

Among 80 paired observations, 12→24-week transitions were 0→0: 13, 0→1: 14, 1→0: 0, 1→1: 53. The paired RD in responder rate (24w−12w) was 0.175 (95% CI 0.100–0.263; McNemar p=0.000122). Late response occurred in 14 of 27 (51.9%) 12-week nonresponders. Mean OABSS improved by 3.64±2.25 at 12 weeks and 5.14±2.63 at 24 weeks.

Conclusions

Extending TTNS to 24 weeks was associated with additional symptom improvement and a substantial proportion of late responders without loss of response. Findings suggest that continuing TTNS beyond 12 weeks may be considered for early nonresponders, pending confirmation in randomized, sham-controlled trials.

Supplementary Information

The online version contains supplementary material available at 10.1007/s00345-026-06249-9.

Introduction

Pediatric overactive bladder (POAB) is a form of lower urinary tract dysfunction. According to the International Children’s Continence Society (ICCS), overactive bladder (OAB) is defined as urgency, with or without urinary incontinence (UI), usually accompanied by increased voiding frequency (VF) and nocturia, in the absence of urinary tract infection (UTI) or other evident organic pathology [1, 2]. Among children aged 5–10 years, the prevalence is 5–12%, whereas among older adolescents (16–18 years) it is approximately 0.5%. Published data indicate that about one third of children with OAB may continue to experience similar symptoms into adulthood. A Swedish study reported a prevalence of roughly 7% in girls and 3.8% in boys [3, 4].Pediatric lower urinary tract symptoms (LUTS) are associated with nocturnal enuresis, recurrent urinary tract infections (UTIs), and constipation [5].Lower urinary tract symptoms (LUTS) are also associated with psychological changes and social embarrassment, including school-based social difficulties and problems with interactions both within and outside the family; related behavioral disorders and social withdrawal may also occur. These issues often improve after successful treatment of urinary incontinence [6].

In recent years, the incidence of OAB has shown a steady upward trend, severely compromising patients’ social participation and significantly reducing health-related quality of life [4, 7].Currently, management of OAB primarily includes behavioral therapies (e.g., bladder training, pelvic floor muscle training, biofeedback), oral anticholinergic agents, β3-adrenergic receptor agonists, and intradetrusor injections of onabotulinumtoxinA [1, 2, 7, 8].Despite the availability of bladder training and pharmacotherapy, many patients fail to achieve satisfactory symptom control, underscoring the need for new treatment options that are cost-effective and effective while being well tolerated for long-term use [9–11].

Transcutaneous tibial nerve stimulation (TTNS) is a peripheral neuromodulation technique. Early international studies have shown that non-implantable, non-invasive TTNS can effectively alleviate symptoms of overactive bladder (OAB) and improve urodynamic parameters [12, 13].In recent years, tibial nerve stimulation (TNS), a noninvasive neuromodulation technique, has shown promising efficacy in adults with OAB; however, systematic clinical studies in children remain scarce. As a specific form of TNS, transcutaneous tibial nerve stimulation (TTNS) offers practical advantages, including ease of operation, wearability, and good treatment adherence [3, 9, 11].Multiple international studies have adopted 12 weeks (3 months) as the initial assessment time point and 24 weeks (6 months) as the observation window for evaluating maintenance of efficacy and potential further improvement [3, 9].Evidence on the influence of treatment duration on therapeutic efficacy in pediatric populations remains limited.

Within stepwise pediatric OAB care, structured urotherapy (education, timed voiding, fluid optimization, and bowel management) remains the first-line approach and can achieve meaningful improvement in a substantial proportion of children (often reported around 40–70%). Neuromodulation, including TTNS, is generally positioned as a second-line (adjunct) option for children with persistent symptoms after urotherapy with or without pharmacotherapy, when appropriate. We adopted a 24-week regimen to evaluate whether extending TTNS beyond the conventional 12-week course provides incremental benefit and to characterize “late responders” who improve after an initial non-response [14–16].

Overactive bladder (OAB) in children is common and burdensome. Transcutaneous tibial nerve stimulation (TTNS) is a noninvasive neuromodulation option with growing evidence in pediatric lower urinary tract dysfunction. Whether extending treatment beyond 12 weeks adds meaningful benefit—particularly in early nonresponders—remains clinically important but under-explored. We conducted an estimation-focused prospective study to quantify the added benefit of extending TTNS to 24 weeks and to characterize the profile of late responders.

Materials and methods

Study design and participants

Prospective, single-arm interventional cohort conducted at Qilu Hospital, Jinan, Shandong Province, China. Eligible children had OAB by standardized criteria and received home-based TTNS with scheduled assessments at baseline, 12 weeks, and 24 weeks. Parental consent and child assent were obtained as applicable. The study adhered to institutional ethics approval (approval No. KYLL-202502(YJ)−004). Eligibility criteria (inclusion and exclusion) are provided in Supplementary Table 1. Children aged 6 to 18 years (either sex) were eligible. Participants were required to have primary pediatric overactive bladder (POAB) or POAB symptoms, meeting at least one of the following clinical features: (i) urgency, defined as a sudden and compelling desire to void that is difficult to defer; (ii) urgency urinary incontinence, defined as involuntary leakage occurring with urgency or immediately after urgency; or (iii) urinary frequency, defined as daytime voiding frequency ≥ 10 times/day with nighttime voiding ≥ 2 times/night and a voided volume < 200 mL per void in children aged 6–11 years, or daytime voiding frequency ≥ 8 times/day with nighttime voiding ≥ 2 times/night and a voided volume < 200 mL per void in children aged ≥ 12 years.

Table 1.

Outcome summary (OABSS and responders)

Measure	N	Mean ± SD/Rate	Mean/Rate (95% CI)
Baseline OABSS	80	10.38 ± 2.35	9.85–10.90
12-week OABSS	80	6.74 ± 2.85	6.10–7.37
24-week OABSS	80	5.24 ± 2.87	4.60–5.88
Responder rate (12w, MCID ≥ 3)	80	66.3%	55.4%–75.7%
Responder rate (24w, MCID ≥ 3)	80	83.8%	74.2%–90.3%

The diagnosis of POAB was confirmed using the Overactive Bladder Symptom Score (OABSS; total score range 0–15). Children were considered eligible if the urgency item score was ≥ 2 and the total OABSS was ≥ 3, with symptoms persisting for ≥ 3 months. Additional inclusion criteria were the ability to use the toilet independently, and provision of written informed consent by the child’s parent/guardian (voluntary participation).

To minimize pharmacologic confounding, participants were required to be free from OAB pharmacotherapy (e.g., anticholinergics or β3-agonists) for at least 4 weeks before baseline (washout period), and initiation of any OAB medications was not permitted during the 24-week study; any need for rescue treatment was documented as a protocol deviation and considered in sensitivity analyses.

Prior exposure to OAB medications and prior neuromodulation for lower urinary tract symptoms (e.g., PTNS/TTNS/TENS) were assessed and recorded at enrollment; notably, none of the enrolled children reported a prior history of tibial nerve neuromodulation. Therefore, prior neuromodulation was unlikely to confound treatment response in this cohort.

Because bowel management is a component of structured urotherapy, constipation treatments were, in principle, required to have been stable for ≥ 4 weeks before baseline and were intended to remain unchanged during follow-up; any dose change or new initiation was recorded.

At enrollment, symptom duration and prior conservative management were documented. Consistent with stepwise pediatric OAB care, most participants had received behavioral urotherapy in routine practice, and a subset had previously tried OAB pharmacotherapy with insufficient symptom control; TTNS was therefore offered as a second-line adjunct in a cohort with a relatively long symptom duration.

Children were excluded if they had stress urinary incontinence, congenital urinary tract malformations causing persistent (unresolved) lower urinary tract symptoms, neurological disease or injury with persistent (unresolved) lower urinary tract symptoms, or had undergone lower urinary tract surgery within the past 3 months. Other exclusion criteria included serious comorbidities (e.g., malignancy) that could affect health status or trial outcomes; skin breakdown, malignancy, or acute suppurative inflammation at the planned plantar electrode site; implanted therapeutic electronic devices (e.g., pacemaker, implantable cardioverter–defibrillator, or sacral neuromodulation device); unresolved urinary tract infection; urinary tract stones causing lower urinary tract symptoms; significant voiding obstruction with post-void residual urine > 100 mL; or any other condition deemed inappropriate for participation by the investigators. The research flow chart is depicted in Fig. 1.

Fig. 1.

Participant Flow

Outcomes and MCID

The primary endpoint was change in OABSS from baseline to 12 and 24 weeks, with responder status defined a priori as OABSS reduction ≥ 3 points (MCID). Secondary outcomes included voiding-diary metrics and PedsQL (MCID ≈ 4.5 points). Instrument availability. OABSS and PedsQL are copyrighted instruments; therefore, the full English item text is not reproduced in this manuscript or the Supplementary Materials. Official English versions can be obtained from the instrument developers/publishers and can be provided to the editorial office upon request. In the Supplementary Materials, we provide an English-language description of the domains, response options, and scoring procedures to facilitate interpretation.

Sample size and precision

Sample size/precision (post hoc): The paired RD in responder rate (24w−12w) was 0.175 with 95% bootstrap CI 0.100 to 0.263; McNemar’s exact p = 0.000122. Because this is a single-arm study without a parallel control group, RD is presented descriptively. No a priori sample-size or power calculation was performed; analyses are estimation-focused.

Data collection

Uroflowmetry and invasive urodynamic testing were not protocol-mandated in this cohort and were not included as study outcomes; if such tests were performed for routine clinical reasons, the results were not analyzed in the present report. Accordingly, participants were required to maintain a daily 24-hour voiding diary throughout the treatment period and to complete the Overactive Bladder Symptom Score (OABSS) questionnaire and the Pediatric Quality of Life Inventory (PedsQL 4.0) at baseline, at 3 months, and at 6 months [9, 12, 17, 18].

Symptom severity was assessed using the Overactive Bladder Symptom Score (OABSS), a 4-item instrument covering daytime frequency (0–2), nighttime frequency/nocturia (0–3), urgency (0–5), and urgency urinary incontinence (0–5), yielding a total score of 0–15 (higher scores indicate more severe symptoms) [19, 20]. At baseline, OAB was operationally defined as an urgency item score ≥ 2 together with a total OABSS ≥ 3, and symptom severity was categorized as mild (3–5), moderate (6–11), or severe (12–15). A clinically meaningful response was prespecified as a ≥ 3-point decrease in total OABSS. Symptom terminology and definitions followed International Continence Society (ICS) recommendations; eligibility and clinical diagnosis were determined by symptom-based assessment consistent with ICS terminology rather than by a single OABSS cutoff alone.

Health-related quality of life (HRQoL) was assessed using the PedsQL 4.0 Generic Core Scales (parent-proxy version), which includes 23 items across physical, emotional, social, and school functioning. Items are rated on a 5-point Likert scale (0–4), reverse-scored, and linearly transformed to a 0–100 scale, with higher scores indicating better HRQoL; changes were interpreted against a commonly used minimal clinically important difference (MCID) of ~ 4–5 points. PedsQL was chosen to capture broad functioning and well-being beyond urinary symptoms and to allow comparability with other pediatric cohorts using a validated parent-proxy format. As a generic HRQoL instrument, PedsQL supports cross-condition and cross-study comparability; disease-specific tools such as PINQ can be added in future studies to better capture urinary-specific burden. Disease-specific measures (e.g., the Pediatric Incontinence Questionnaire [PINQ]) may offer greater urinary-focused sensitivity and should be considered in future work [21].

Ethics

The study was approved by the Institutional Review Board of Qilu Hospital, Shandong University (approval No. KYLL-202502(YJ)−004). Written informed consent was obtained from all parents/guardians, and written assent was obtained from the children. The trial was conducted in accordance with the Declaration of Helsinki and Good Clinical Practice guidelines [12].

Treatment

TTNS was delivered using a portable home-based stimulator (TTNS-W1). Surface electrodes were placed over the posterior tibial nerve at the ankle. An overview of the TTNS-W1 device and a schematic of standardized electrode placement are provided in Supplementary Fig. 4, and representative photographs of home-based use are shown in Supplementary Fig. 3.Each session lasted 30 min and was performed every other day for 24 weeks. The pulse width was fixed at 200 µs (0.2 ms). Stimulation began at 10 Hz; for participants who did not meet the prespecified response criterion at each 4-week assessment, the frequency was increased by 1–2 Hz (up to 30 Hz) while other settings were unchanged. Current intensity was titrated to a comfortable sensory level (typically 10–30% above sensory threshold) and was always kept below the motor threshold.

Concomitant treatments: All participants received standardized urotherapy as background care throughout follow-up, including pediatric fluid-intake guidance (individualized by body weight and activity level), avoidance of caffeinated beverages, and timed voiding training [12]. Constipation was screened at enrollment and, when present, was addressed with standardized behavioral counseling (regular toileting, increased water/fiber intake, and avoidance of stool withholding); glycerin suppositories/enemas were permitted as needed, without routine use of pharmacologic laxatives. No OAB pharmacotherapy or additional neuromodulation was allowed or used during the 24-week study period.

Health education and psychological support: A pediatric urology specialist contacted families at the end of each week via WeChat to review the child’s 24-hour voiding diary, reinforce urotherapy instructions, and provide psychological counseling to the child and caregivers as needed [12].

To maximize adherence and retention, caregivers received standardized training at baseline (electrode placement, device operation, and safety checks), together with an illustrated written guide. Families recorded each stimulation session in a home log, and study staff used scheduled WeChat contacts to verify session completion, troubleshoot device use, and collect adverse-event information. Missed sessions triggered reminder messages and make-up scheduling when feasible, which likely contributed to the high completion rate (80/80).

The planned regimen comprised a 12-week core course plus a 12-week extension (total 24 weeks). The device’s ability to export detailed treatment logs enabled the use of planned/completed sessions and cumulative stimulation duration as the primary metrics for evaluating adherence and treatment dosage. An overview of the TTNS device and illustrative photographs of patient use (including standardized electrode placement) are provided in Supplementary Figs. 3,4.

Fig. 3.

Responder transition counts (12w→24w)

Fig. 4.

Response Regimen Evolution: 12→24 Weeks

Outcome measures and stratification

Assessment Metrics: ①Baseline characteristics: age, sex, race/ethnicity, symptom duration, body mass index (BMI).②24-h voiding diary: daytime voiding frequency, nighttime voiding frequency, episodes of urgency, episodes of urgency urinary incontinence (UUI), and daily fluid intake.③Overactive Bladder Symptom Score (OABSS): The validated Chinese-language OABSS was administered at baseline, week 12, and week 24 to obtain total and item scores (daytime frequency, nighttime frequency, urgency, and UUI). Daytime/nighttime frequencies and urgency/UUI counts within the OABSS were calculated from the mean of the 24-h voiding diary over the 3 days preceding questionnaire completion at each time point. Item scoring followed the standard anchors (0–2, 0–3, or 0–5), total range 0–15, with higher scores indicating more severe symptoms.The primary focus was change in OABSS total score.④Pediatric Quality of Life Inventory (PedsQL 4.0): Parent-proxy version at baseline, week 12, and week 24; total score and four domains (physical, emotional, social, and school functioning). All domain and total scores were transformed to a 0–100 scale, with higher scores indicating better health-related quality of life (HRQoL) [18]。⑤Adverse events (AEs): All AEs occurring during treatment were recorded [12].

Responder definition: Therapeutic response was defined a priori as a ≥ 3-point decrease in OABSS total score from baseline at week 12; otherwise, participants were categorized as non-responders [12, 17].This threshold, informed by prior literature and clinical judgment, corresponds to approximately 30% improvement relative to the mean baseline OABSS and serves as one minimal clinically important difference (MCID) criterion [22]. The same criterion was applied at week 24, and response status was compared with week 12 to assess transitions.

Primary Efficacy Endpoints: Primary efficacy endpoints were the change in total Overactive Bladder Symptom Score (OABSS) from baseline to weeks 12 and 24 and the responder rates at weeks 12 and 24. The principal focus of this study was change in the OABSS total score [12]. To specifically evaluate the benefit of extending treatment beyond the standard 12-week assessment point, we evaluated the within-participant change in responder status from week 12 to week 24 and reported the paired risk difference (RD) in responder rate together with a 95% bootstrap confidence interval as a descriptive measure of the absolute increase in response with extension.

Secondary Endpoints: (1) Changes in voiding-diary metrics: including the 24-hour average number of voids (tabulated separately for daytime and nighttime), average daily urgency episodes, and average daily urgency urinary incontinence (UUI) episodes. Parents were instructed to assist children in recording daily fluid intake, the time and volume of each void, and the occurrence of urgency or incontinence episodes. At each evaluation timepoint (baseline, week 12, and week 24), three consecutive days of voiding diaries were completed, and changes from baseline in the aforementioned metrics were calculated.(2) Changes in health-related quality of life (HRQoL): Assessed using the parent-proxy report version of the Pediatric Quality of Life Inventory (PedsQL) 4.0 Generic Core Scales. This instrument encompasses four domains: physical functioning, emotional functioning, social functioning, and school functioning. Higher total and domain scores indicate better HRQoL. PedsQL total and domain scores were recorded at baseline, week 12, and week 24. Changes from baseline were compared at each follow-up assessment.(3) Transition-matrix analysis: Response status (responder/non-responder) at week 12 and week 24 was compared. This analysis aimed to construct a transition matrix and visualize population flows using Sankey diagrams (e.g., depicting the proportion of week-12 non-responders who converted to responders by week 24).(4) Attainment of minimal clinically important difference (MCID): In addition to the pre-specified OABSS-based MCID (response defined as OABSS decrease ≥ 3 points), MCID thresholds for objective voiding-diary metrics and HRQoL were defined a priori based on prior literature and scale properties. For objective voiding-diary endpoints (number of voids, urgency episodes, UUI episodes), MCID attainment was defined as a ≥ 50% reduction from baseline in the relevant frequency. For the PedsQL total score, MCID attainment was defined as an increase of ≥ 4.5 points from baseline (as per the PedsQL manual, 4.5 points approximates the minimal threshold for a perceptible change in children). The proportions of patients achieving MCID for each metric at weeks 12 and 24 were calculated and compared using paired statistical tests.(5)Subgroup analyses: Subgroups were defined based on baseline characteristics (e.g., sex, age, body mass index [BMI], race/ethnicity). Week-24 treatment response outcomes were analyzed within these subgroups. Responder rates and the magnitude of therapeutic benefit derived from treatment extension were estimated for each subgroup. Binary logistic regression models (for responder status) and Cox proportional hazards regression models (assessing time-to-response among initial non-responders) were employed to explore potential predictors of response. Forest plots were generated to visualize effect sizes across the different subgroups.

Safety assessment

Participants were contacted weekly by telephone to solicit any discomfort or AEs, and detailed interviews and physical examinations were performed during clinic visits. For each AE, the onset date, duration, manifestations, severity, management, and outcome, as well as causality in relation to TTNS, were recorded [23–25]. Particular attention was given to local cutaneous reactions to electrical stimulation, pain/discomfort, and other systemic adverse effects. All AEs were coded using the Medical Dictionary for Regulatory Activities (MedDRA), and two investigators independently adjudicated relatedness to TTNS; disagreements were resolved by a third-party expert [26, 27]. AE incidence within the first 12 weeks and within 24 weeks was summarized, and AEs were compared between week-12 responders and non-responders to assess potential bias related to differential efficacy.

Statistical analysis

Analyses were conducted in SPSS 25.0 and R 4.1 (including ggplot2); figures were additionally prepared in GraphPad Prism 9. All tests were two-sided with α = 0.05. P values were reported to three decimals; values < 0.001 were shown as “<0.001”. Normality was assessed by Shapiro–Wilk and homogeneity of variance by Levene. Approximately normal continuous variables are summarized as mean ± SD and compared with independent-samples t tests (or paired t tests for within-subject contrasts); non-normal or ordinal data are presented as median (IQR) and compared with the Mann–Whitney U or Wilcoxon signed-rank test, with the Hodges–Lehmann median difference and 95% CI where appropriate. For multiple groups, we used the Kruskal–Wallis test with Bonferroni-adjusted pairwise comparisons.

For outcomes measured at multiple time points, overall differences were first tested by the Friedman test; if significant, Bonferroni-adjusted pairwise comparisons were performed (α′ = 0.017 for three contrasts). To better exploit continuous data and adjust for baseline imbalances, we fitted linear mixed-effects models (LMMs) and generalized estimating equations (GEEs). Fixed effects included time (12 vs. 24 weeks), 12-week responder status (responder vs. non-responder), and their interaction (time×responder); the baseline value of each outcome was included as a covariate, with a random intercept for subjects. When LMM convergence or distributional assumptions were not satisfied, GEEs with an exchangeable correlation structure and robust (sandwich) SEs were used as sensitivity analyses. For markedly skewed count outcomes (e.g., episode counts), we ran verification analyses within the GEE framework using log links with Poisson or negative binomial distributions.

Categorical variables are reported as n (%) and compared using χ² or Fisher’s exact tests; paired proportions (e.g., MCID attainment; responder rates at 12 vs. 24 weeks) were compared with McNemar’s test. 95% CIs for proportions used the Wilson method; CIs for risk difference (RD) and odds ratio (OR) were obtained via percentile bootstrap with 1,000 resamples. Agreement was assessed with Cohen’s κ and Gwet’s AC1 (both with 95% CIs), noting AC1 as a complementary metric when prevalence imbalance may bias κ.

Primary inferences were based on the full analysis set (FAS; intention-to-treat, n = 80); the per-protocol set (PPS) repeated primary analyses as sensitivity. Missing data were handled by last observation carried forward (LOCF) for the primary endpoint and safety outcomes; other secondary outcomes were analyzed on available cases as exploratory. Unless otherwise specified, we report effect sizes with 95% CIs (e.g., mean difference/Cohen’s d or median difference/Hodges–Lehmann) and apply Bonferroni only to pre-specified pairwise multiple comparisons; remaining secondary analyses are exploratory without further α-level adjustment.

Results

Primary outcome

OABSS decreased from 10.38 ± 2.35 at baseline to 6.74 ± 2.85 at 12 weeks and to 5.24 ± 2.87 at 24 weeks (Δ = 5.14 ± 2.63)(Table 1; Fig. 2). The mean decrease from baseline was 3.64 ± 2.25 points at 12 weeks (95% bootstrap CI 3.15–4.13) and 5.14 ± 2.63 points at 24 weeks (95% bootstrap CI 4.56–5.71) (Table 1; Fig. 2). The mean additional decrease from 12 to 24 weeks was 1.50 points (95% bootstrap CI 1.23–1.79). Key outcomes are summarized in Table 5; the complete standard reporting-format table is provided in Supplementary Table 8.

Fig. 2.

Mean OABSS with 95% confidence intervals at baseline, 12 weeks, and 24 weeks

Table 5.

Key outcomes (absolute change) overall and stratified by week-12 responder status

Endpoint	Metric	Overall (n=80, FAS)Mean±SD	12-week Responders (n=53)Mean±SD	12-week Non-responders (n=27)Mean±SD	Between-group P(Non-responder vs Responder)	Paired P in Non-responders(3→6 months)
Primary outcome
OABSS total score	Δ at 3 months (absolute)	3.64 ± 2.25	4.94 ± 1.49	1.07 ± 0.83	NA
OABSS total score	Δ at 6 months (absolute)	5.14 ± 2.63	6.51 ± 1.98	2.44 ± 1.37	<0.001
OABSS total score	Δ from 3→6 months (absolute)	1.52 ± 1.24	1.60 ± 1.32	1.37 ± 1.08	0.555	<0.001
Secondary outcomes – Voiding diary
daytime frequency (per day)	Δ at 3 months (absolute)	3.65 ± 3.00	4.12 ± 3.15	2.74 ± 2.50	0.053
daytime frequency (per day)	Δ at 6 months (absolute)	5.23 ± 3.90	5.68 ± 4.04	4.35 ± 3.50	0.203
daytime frequency (per day)	Δ from 3→6 months (absolute)	1.77 ± 2.11	1.77 ± 1.99	1.75 ± 2.37	0.666	<0.001
nighttime frequency (per night)	Δ at 3 months (absolute)	1.11 ± 0.81	1.32 ± 0.78	0.70 ± 0.71	0.001
nighttime frequency (per night)	Δ at 6 months (absolute)	1.20 ± 0.87	1.35 ± 0.82	0.90 ± 0.91	0.029
nighttime frequency (per night)	Δ from 3→6 months (absolute)	0.19 ± 0.42	0.14 ± 0.26	0.30 ± 0.62	0.358	0.140
daily urgency episodes	Δ at 3 months (absolute)	1.79 ± 1.81	2.01 ± 1.82	1.36 ± 1.74	0.090
daily urgency episodes	Δ at 6 months (absolute)	2.23 ± 2.10	2.38 ± 2.29	1.91 ± 1.68	0.540
daily urgency episodes	Δ from 3→6 months (absolute)	0.55 ± 0.77	0.50 ± 0.85	0.67 ± 0.58	0.036	0.002
daily urge incontinence episodes	Δ at 3 months (absolute)	1.99 ± 2.03	2.53 ± 2.22	0.91 ± 0.97	<0.001
daily urge incontinence episodes	Δ at 6 months (absolute)	2.32 ± 2.11	2.74 ± 2.30	1.48 ± 1.35	0.020
daily urge incontinence episodes	Δ from 3→6 months (absolute)	0.42 ± 0.56	0.31 ± 0.49	0.62 ± 0.63	0.004	<0.001
Secondary outcomes – PedsQL
PedsQL Total score	Δ at 3 months (absolute)	7.92 ± 5.60	8.80 ± 5.56	6.20 ± 5.38	0.075
PedsQL Total score	Δ at 6 months (absolute)	9.08 ± 6.47	10.21 ± 6.55	6.84 ± 5.79	0.052
PedsQL Total score	Δ from 3→6 months (absolute)	1.41 ± 1.63	1.66 ± 1.74	0.93 ± 1.27	0.083	0.049

Note: For OABSS and voiding-diary outcomes, positive Δ indicates improvement (Δ = Baseline - Follow-up). For PedsQL, positive Δ indicates improvement (Δ = Follow-up - Baseline).

Responder transitions and paired RD

Responder transitions from week 12 to week 24 were: 0→0 (n = 13), 0→1 (n = 14), 1→0 (n = 0), and 1→1 (n = 53). The paired risk difference (RD) in responder rate (week 24 minus week 12) was 0.175 (95% CI 0.100–0.263; McNemar P = 0.000122), indicating an absolute 17.5% increase in responders with treatment extension. Among 12-week nonresponders, 14/27 (51.9%) achieved a late response, and no loss of response was observed (Tables 2 and 3; Fig. 3).

Table 2.

Responder transition matrix (12w→24w)

12w non-responder (0)	24w non-responder (0)	24w responder (1)	Row total
	13	14	27
12w responder (1)	0	53	53
Total	13	67	80

Table 3.

Paired RD and precision

Paired RD (24w−12w)	95% CI (bootstrap)	McNemar exact p	Late response 0→1	Loss of response 1→0
0.175	[0.100, 0.263]	0.000122	14	0

Baseline characteristics

A total of 80 children with OAB were included; 53 (66.3%) were responders at week 12 and 27 (33.8%) were non-responders.

Baseline characteristics were comparable between groups: the median age was 8.0 years in both (P = 0.229),(overall range, 6–15 years; responders, 6–15 years; non-responders, 6–14 years; P = 0.229); sex distribution was similar (P = 0.813); BMI was comparable [21.20 ± 1.46 vs. 21.24 ± 1.76 kg/m²; P = 0.915]; duration of symptoms was similar [1.50 (1.20–2.40) vs. 1.70 (1.15–2.25) years; P = 0.787]; and there was no significant difference in race/ethnicity (P = 0.774) (Fig. 4).

Baseline OABSS and PedsQL did not differ between groups [OABSS: 10.0 (9.0–12.0) vs. 10.0 (8.0–12.0), P = 0.629; PedsQL: 84.17 ± 6.00 vs. 83.57 ± 5.14, P = 0.647].

During follow-up, symptom scores diverged markedly: at 3 months the responder group had significantly lower OABSS than non-responders [5.53 ± 2.23 vs. 9.11 ± 2.41; P < 0.001]; at 6 months the difference persisted and remained significant [4.00 (2.00–5.00) vs. 7.00 (5.50–10.00); P < 0.001].

With respect to quality of life, at 3 months PedsQL scores were higher in responders with a near-significant difference [94.57 (90.22–96.74) vs. 90.22 (84.24–95.65); P = 0.063], and by 6 months the difference reached statistical significance [95.65 (92.39–98.91) vs. 91.30 (85.87–97.28); P = 0.023].

Overall, given balanced baselines, responders exhibited greater symptom relief at both 3 and 6 months and demonstrated superior health-related quality of life at 6 months.

These findings indicate no significant pre-intervention differences between responders and non-responders, supporting their comparability(Table 4).

Table 4.

Baseline characteristics

Variable	Overall (n = 80)	12-week Responders (n = 53)	12-week Non-responders (n = 27)	P value
Age(years)	8.00 [7.00, 9.00](range 6–15)	8.00 [7.00, 9.00](range 6–15)	8.00 [7.50, 9.00](range 6–14)	0.229
Sex				0.813
Female	36 (45.0%)	23 (43.4%)	13 (48.1%)
Male	44 (55.0%)	30 (56.6%)	14 (51.9%)
BMI	21.22 ± 1.56	21.20 ± 1.46	21.24 ± 1.76	0.915
Symptom duration, years	1.55 [1.20, 2.40]	1.50 [1.20, 2.40]	1.70 [1.15, 2.25]	0.787
Ethnicity				0.774
Hui	5 (6.2%)	3 (5.7%)	2 (7.4%)
Han	73 (91.2%)	48 (90.6%)	25 (92.6%)
Uyghur	1 (1.2%)	1 (1.9%)	0 (0.0%)
Mongolian	1 (1.2%)	1 (1.9%)	0 (0.0%)
OABSS score
Baseline OABSS	10.00 [8.75, 12.00]	10.00 [9.00, 12.00]	10.00 [8.00, 12.00]	0.629
OABSS at 3 months	6.74 ± 2.85	5.53 ± 2.23	9.11 ± 2.41
OABSS at 6 months	4.50 [3.00, 7.00]	4.00 [2.00, 5.00]	7.00 [5.50, 10.00]
PedsQL score
Baseline PedsQL	83.97 ± 5.70	84.17 ± 6.00	83.57 ± 5.14	0.647
PedsQL at 3 months	92.39 [88.86, 96.74]	94.57 [90.22, 96.74]	90.22 [84.24, 95.65]	0.063
PedsQL at 6 months	95.65 [88.86, 97.83]	95.65 [92.39, 98.91]	91.30 [85.87, 97.28]	0.023

12-Week Efficacy and Effects of Extension to 24 Weeks

Overall response and temporal trend

All participants (n = 80) completed assessments at baseline, week 12, and week 24, with no major protocol deviations; therefore, the full analysis set (FAS) and per-protocol set (PPS) were identical (both n = 80). Global time effects across the three time points were evaluated using the Friedman test; when significant, pairwise Wilcoxon signed-rank tests with Holm adjustment were performed (Supplementary Tables 2–4).

In the FAS, the OABSS-based responder rate increased from 66.3% (53/80) at week 12 to 83.8% (67/80) at week 24. Responder transitions from week 12 to week 24 were 0→0 (n = 13), 0→1 (n = 14), 1→0 (n = 0), and 1→1 (n = 53), corresponding to an absolute paired risk difference (RD) of 0.175 (95% CI 0.100–0.263) and a statistically significant McNemar exact test (P = 0.000122; Supplementary Tables 5–7). Among week-12 non-responders, 51.9% (14/27) converted to responders by week 24; no loss of response was observed.

Consistent with the responder analysis, OABSS declined continuously from baseline to 3 months and 6 months (Fig. 5). From baseline to 6 months, the mean improvement in OABSS total score was 5.14 ± 2.63 points; the additional improvement from 3 to 6 months was 1.52 ± 1.24 points (both P < 0.001 for within-group change; Supplementary Tables 2–4; summary estimates in Table 5). Stratified by week-12 status, week-12 responders achieved greater baseline-to-6-month improvement than non-responders (P < 0.001), whereas the incremental gain during the extension period (3→6 months) was comparable between groups (P = 0.555 for absolute change; Table 5; Figs. 6 and 7).

Fig. 5.

OABSS Temporal Trends (Baseline, Month 3, Month 6) Stratified by 12-Week Response Status (Non-responders vs. Responders)

Fig. 6.

Complete individual-level distribution visualization enabling tail behavior assessment

Fig. 7.

Consistency of the 3→6-month incremental benefit (Δ) across major subgroups (forest plot) Note: The x-axis denotes the direction of improvement (rightward = greater improvement).

Secondary outcomes and health-related quality of life

Diary-based outcomes showed continued benefit with treatment extension (Table 5; percentage changes are summarized in Supplementary Table 8). Daytime voiding frequency continued to decrease from 3 to 6 months (mean reduction 1.77 ± 2.11 episodes/day; within week-12 non-responders P < 0.001), whereas the principal improvement in nighttime voiding frequency occurred during the first 3 months (3→6 months within week-12 non-responders P = 0.140; Table 5). Urgency and urgency urinary incontinence (UUI) were both lower in week-12 responders at 3 and 6 months; however, week-12 non-responders accrued greater additional benefit during months 3–6 (urgency: between-group P = 0.036; UUI: P = 0.004; Table 5), and percentage-based analyses were concordant (both P ≤ 0.023; Supplementary Table 8).

Health-related quality of life improved in parallel with symptom relief. The PedsQL total score increased with treatment, with a modest additional gain from 3 to 6 months (+ 1.41 ± 1.63; within week-12 non-responders P = 0.049; Table 5). Domain-level analyses are presented in Supplementary Table 8: social and school/role functioning showed mild additional improvement in responders (P = 0.036 and 0.046, respectively), whereas the physical domain was not significant.

Individual heterogeneity

To characterize individual-level heterogeneity during the extension period, cumulative distribution function (CDF) plots of Δ(3→6 months) demonstrated a rightward shift for most endpoints except nighttime voiding, indicating that most children experienced additional improvement with continued treatment; the shift was more pronounced for urgency and UUI among week-12 non-responders (Fig. 6).

Exploratory dose-outcome association

Exploratory exposure–outcome analyses suggested a nonlinear association between 24-week OABSS improvement and stimulation exposure in spline models adjusted for baseline OABSS (Supplementary Fig. 5). In quartile analyses of stimulation current, improvements were broadly comparable across exposure bins without a clear monotonic trend (also shown in Supplementary Fig. 5). Supplementary Fig. 6 presents the quartile-based analyses of stimulation current, showing broadly comparable improvements across bins without a clear monotonic trend. These findings are hypothesis-generating and should be interpreted cautiously.

Adherence and exposure (dose)

All 80 participants completed the scheduled assessments at baseline, week 12, and week 24, with no treatment discontinuations and no major protocol deviations. Home TTNS exposure was monitored using caregiver session logs in combination with weekly staff contacts. Adherence/dose summaries are provided in Table 7, and the distributions of completed sessions and stimulation current are shown in Figs. 8 and 9, respectively. Stimulation current had a median of 10.5 mA (IQR 9.5–12.0), a mean of 10.93 ± 1.94 mA, and a range of 8.0–19.0 mA (n = 80). Overall, exposure appeared consistent across participants and adequate to support the efficacy analyses(Table 7; Figs. 8 and 9).

Table 7.

Adherence and dose summary

Adherence/Dose	N	Median [IQR]	Mean ± SD
Session-completion category (0–2)	80	1.00 [1.00, 2.00]	1.44 ± 0.57
stimulating current	80	10.50 [9.50, 12.00]	10.93 ± 1.94

Note: Session-completion category was derived from planned vs completed sessions (0 = did not complete planned sessions; 1 = completed planned sessions; 2 = completed planned sessions with make-up/additional sessions).

Fig. 8.

Distribution of session-completion category (0–2)

Fig. 9.

Distribution of stimulation current

Subgroup analysis

Across the prespecified subgroups (sex, age strata [6–12 years/13–18 years], BMI-for-age Z score per WHO 2007 [normal/overweight/obesity], symptom-duration tertiles, and race/ethnicity), effect consistency was evaluated using multivariable logistic regression (24-week response) and Cox models (time-to-response). The adjusted ORs/HRs in each subgroup were close to 1.0, with 95% CIs generally crossing 1; both model types were directionally consistent, and no statistically or clinically meaningful heterogeneity was observed. These findings suggest uniform treatment efficacy across baseline strata, with no evident effect modification (full effect sizes and intervals are provided in Supplementary Fig. 1–2 and the subgroup forest plots).

Adverse events and safety

Over 24 weeks, 11 AE episodes were recorded among 10 children (12.5%), with no serious adverse events, no treatment discontinuations, and no irreversible reactions. The only device-related AE was mild patch-site pruritus/rash in 5/80 (6.2%), typically emerging 2–3 weeks after initiation; presentations were limited to mild local erythema/pruritus without vesicles or marked inflammation and resolved within several days after cleansing, relocating the electrode, or brief interruption/down-titration. Incidence in responders vs. non-responders was 5.7% vs. 7.4% (P = 1.00). Other events were adjudicated unrelated to TTNS: common cold (n = 6), pneumonia (n = 2), varicella (n = 1), catheter-associated UTI (n = 1), and traumatic fracture (n = 1). Notably, the catheter-associated UTI occurred after short-term urethral catheterization performed for an intercurrent condition/procedure (unrelated to TTNS); the episode resolved with standard care and was adjudicated as not device-related. Rates did not differ significantly between groups, and all cases improved with standard care. Vital signs remained stable throughout treatment, no neurologic injury signs were detected, children reported no pain or intolerability, and treatment courses were completed uneventfully; no cumulative risk was observed with extension of therapy (Table 6).

Table 6.

Adverse events by 12-week response status

Adverse event	12-week responders (n=53)	12-week non-responders (n=27)	Total (n=80)	Between-group comparison
Patch-site skin pruritus	3/53 (5.7%)	2/27 (7.4%)	5/80 (6.2%)	P = 1.00
Upper respiratory tract infection	4/53 (7.5%)	2/27 (7.4%)	6/80 (7.5%)	—
Pneumonia	2/53 (3.8%)	0	2/80 (2.5%)	—
Varicella	1/53 (1.9%)	0	1/80 (1.2%)	—
Urinary tract infection (catheter-associated)	0	1/27 (3.7%)	1/80 (1.2%)	—
Fracture	0	1/27 (3.7%)	1/80 (1.2%)	—

Note: Values are presented as number of cases/total in group (percentage). “Device-related” denotes adverse events judged possibly or probably related to electrical stimulation; the remaining events were intercurrent illnesses or accidents adjudicated as unrelated to treatment. Between-group comparisons used Fisher’s exact test (applicable only to device-related skin pruritus; other events were not compared owing to small numbers).

Discussion

An estimation-focused analysis indicates that extending TTNS from 12 to 24 weeks is associated with additional symptom improvement and a substantial proportion of late responders, without evidence of response loss. These findings argue for continued TTNS beyond 12 weeks in early nonresponders while acknowledging that causal confirmation requires controlled trials.

Under balanced baselines (age, sex, BMI, disease course, race/ethnicity, and OABSS/PedsQL all without between-group differences), this study systematically evaluated the efficacy, safety, and population heterogeneity of TTNS at two time points—12 and 24 weeks. Three principal findings emerged. First, symptomatic benefit was substantial and deepened with treatment extension. Across the cohort, the 6-month reduction in OABSS versus baseline was 5.14 ± 2.63 points (≈ 50%). Using the prespecified MCID (≥ 3 points), the responder rate increased from 66.3% at 12 weeks to 83.8% at 24 weeks, an absolute gain of 17.5% from week 12 to week 24. Notably, 51.9% of children who were non-responders at 12 weeks converted to responders with extended treatment, and no “loss of response” was observed, consistent with a delayed response phenotype and suggesting that continuation beyond 12 weeks may be considered for early non-responders, while acknowledging that causal attribution requires controlled trials. Second, symptom domains exhibited distinct temporal trajectories: daytime-related indices (daytime voiding, urgency, and urgency urinary incontinence) showed statistically significant additional improvement from 3 to 6 months, whereas nighttime voiding improved mainly within the first 12 weeks with limited incremental gain thereafter. This pattern accords with clinical experience and implies that, after 12 weeks, therapeutic strategy may prioritize consolidation and intensification of daytime symptom control. Third, health-related quality of life (PedsQL) improved overall with treatment; by 6 months, between-group differences were significant and favored responders. However, the incremental effect from 3 to 6 months was smaller in magnitude than symptom improvement, suggesting that quality-of-life gains may lag behind symptomatic relief or be constrained by a ceiling effect. We also found that a 6-month course was safe and well tolerated, with no cumulative toxicity or other serious adverse events observed [28].

Under balanced baselines (age, sex, BMI, disease course, and baseline symptom severity), extending TTNS from 12 to 24 weeks was associated with additional symptom improvement and a higher proportion of responders. Across the cohort, mean OABSS decreased by 5.14 ± 2.63 points by 6 months (≈ 50% reduction from baseline). Using the prespecified MCID (≥ 3-point decrease), the responder rate increased from 66.3% at 12 weeks to 83.8% at 24 weeks, corresponding to an absolute increase of 17.5% (bootstrap 95% CI 10.0% to 26.3%; McNemar p = 0.000122). Importantly, 14 children (17.5%) who were non-responders at 12 weeks became responders by 24 weeks (late responders), while no loss of response was observed among initial responders.

Implications

Most pediatric TTNS studies to date have used short courses (8–12 weeks) with responder rates around 60–70%[10]. Using a prespecified and relatively stringent OABSS responder definition (> = 3-point reduction), our study achieved a comparable 12-week effect and further showed that extending home-based TTNS to 24 weeks was associated with a higher overall responder rate at 6 months. This pattern is directionally consistent with adult PTNS maintenance experience, in which consolidation of earlier gains is commonly observed. For children who do not meet symptom goals at 12 weeks, continuing TTNS for an additional 12 weeks may therefore be considered, while acknowledging that causal attribution requires sham-controlled confirmation. Within 6 months, only mild patch-site pruritus was observed (6.2%), which was short-lived and self-limited without discontinuations; no accumulation of adverse events was observed with longer treatment. Notably, Bouali et al. reported a 6-month pediatric TTNS experience; compared with that work, our cohort is larger and our analyses explicitly quantify response transitions to identify and describe late responders [3].

Limitations

This study has several limitations. First, the single-center, non-randomized, unblinded design means that expectancy and placebo effects cannot be excluded; placebo/expectancy responses of approximately 30–40% have been reported in pediatric OAB trials and may inflate apparent effects in single-arm studies. Second, the intervention package included structured urotherapy components and behavioral reinforcement (e.g., timed voiding, diary keeping with weekly feedback, and constipation screening with behavioral counseling when present). Without a control group, improvements cannot be attributed to TTNS alone. Third, symptom changes over 24 weeks may partly reflect natural maturation and developmental gains in continence control, which cannot be separated from treatment effects without an untreated or sham comparator. Fourth, constipation improvement was not a prespecified study endpoint and was not systematically assessed; constipation severity was not quantified using a dedicated validated score, precluding a formal evaluation of constipation improvement as an outcome. Additional limitations include reliance on questionnaires/diaries rather than objective urodynamic measures, a modest sample size with multiple comparisons, real-world variability in caregiver implementation and adherence, and follow-up limited to 24 weeks without post-treatment maintenance assessment Table 7.

Future directions

From a clinical perspective, structured urotherapy should remain the foundation of care. TTNS may be considered as an adjunct/second-line option for children who remain symptomatic after urotherapy (and pharmacotherapy when appropriate), in line with contemporary pediatric care pathways. Our data suggest that continuing TTNS beyond 12 weeks may be considered for early nonresponders, but decisions should be individualized and integrated with background behavioral management. Randomized sham-controlled trials and longer follow-up are needed to define the optimal duration, maintenance regimen, and the incremental benefit of TTNS beyond behavioral support alone.

Conclusions

In this prospective single-arm cohort, extending home-based TTNS from 12 to 24 weeks was associated with further improvement in OAB symptoms and an increased proportion of children meeting the prespecified responder definition, including a meaningful subgroup of late responders, with a favorable safety profile. Because the extension phase lacked a control/sham comparator and was accompanied by structured urotherapy components, these findings should be interpreted as estimation-focused and hypothesis-generating rather than causal.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Fangzheng Cheng was the first author and led the study, including study conception and design, data analysis, and drafting of the manuscript. Jian Wang supervised the project, secured funding, and critically revised the manuscript for important intellectual content. The remaining authors contributed to patient recruitment and data collection, implementation and operation of the TTNS protocol and related procedures, data management and software support, and reviewed and approved the final version of the manuscript.

Funding

This work was supported by the Natural Science Foundation of Shandong Province (grant numbers ZR2022MH276)

Data availability

De-identified individual participant data that support the findings of this study are available from the corresponding author upon reasonable request, subject to institutional review board approval and appropriate data-sharing agreements. Analytic code is also available upon reasonable request. Public deposition is not planned because the dataset includes pediatric health information and may contain potentially identifiable data.

Declarations

Conflict of interest

The authors declare no competing interests.

Ethical approval and consent to participate

This study was conducted in accordance with the Declaration of Helsinki and relevant institutional guidelines. The protocol was reviewed and approved by the Ethics Committee of Qilu Hospital, Shandong University (approval No.: KYLL-202502(YJ)-004; approval date: 2025-02-27). Written informed consent was obtained from the parents or legal guardians of all participants; age-appropriate assent was obtained from children aged ≥ 8 years, in line with institutional policy.