This was a single-centre and prospective case-series study performed at Guang’anmen Hospital, China Academy of Chinese Medical Sciences from August 5, 2017, to July 31, 2022. This study was conducted in accordance with the principles of the Declaration of Helsinki and was approved by the Ethics Committee of Guang’anmen Hospital. All participants and their parents or guardians had signed the written informed consent prior to participation. Trial Registration was available at www.chictr.org.cn (Identifier: ChiCTR1800020222).

The inclusion criteria were: (1) being younger than 18 years old, (2) being diagnosed with CUR caused by LMNL (e.g., sacral plexus, cauda equina, or sacral cord lesions) following lumbosacral surgeries, (3) being unable to void and dependent on catheterization including CIC or indwelling catheterization (IC), (4) CUR had lasted for 3 months or longer.

The exclusion criteria were: (1) bladder outlet obstruction uneliminated, (2) urinary system tumors or stones not removed, (3) major psychological disorders, (4) previous implantation with a cardiac pacemaker or SNM, (5) EA treatment duration of less than 1 week.

PVR was measured by urethral catheter output after children’ spontaneous urination attempts without catheterization or assisted bladder emptying (including increasing abdominal pressure or other auxiliary manual methods). For children with CIC, PVR was the average value of 2 measurements. The interval between two measurements should range from 6 h to 3 days.

Bilateral Ciliao (BL32), Zhongliao (BL33), Huiyang (BL35), Shenshu (BL23), and Sanyinjiao (SP6) were inserted with needles. The parameter of needles used at BL32, BL33 and BL35 were 0.30 mm in diameter and 75 mm in length; the needles at BL23 and SP6 were 0.30 mm in diameter and 40 mm in length (Hwato Brand, Suzhou Medical Appliance Factory, China). Bilateral BL32 and BL33 were needled to a depth of 50 mm–70 mm with an angle of 60°–75° inward and downward, into the second and third sacral foramen. Bilateral BL35 were needled to a depth of 50 mm–60 mm with a slightly superolateral direction. Bilateral BL23 and SP6 were needled vertically to a depth of 25 mm–30 mm. The electric stimulators (SDZ-V electroacupuncture apparatus, Suzhou Medical Appliance Factory, China) were transversely connected to bilateral BL32, BL33, and BL35 with a continuous 5-Hz wave (5–10 mA intensity), while connected to bilateral SP6 with a 10 Hz wave (1–2 mA intensity), with an adjustable current intensity to adapt participants’ tolerance. Needles were retained for 30 min during each treatment session. The participants were treated with EA three sessions a week for consecutive 2–16 weeks. In general, PVR less than 5 ml in children was considered normal and not associated with UTI (30). So, when the participant's PVR had reduced to 5 ml within 16 weeks, EA treatment could be terminated depending on the decisions of their parents or guardians and urologists. If the participant's PVR did not reduce to 5 ml, EA treatment was proceeded till the 16th week. Meanwhile, CIC or IC was administered for all participants.

Responders were defined as participants achieving a reduction of 50% or more in PVR from baseline (31, 32). The primary outcome was the response rate at the 12th week. The secondary outcomes were the change in PVR, in the proportion of participants with severe difficulty in urination and in the proportion of participants with stool retention from baseline, as well as the proportion of participants with catheter removal and the Patient Global Impression of Improvement (PGI-I) score of 1 or 2 at the 12th week. In addition, the proportion of participants without catheterization, the incidence of recurrent symptomatic UTI, hydroureter, and hydronephrosis were assessed in the 24-week follow-up.

Difficulty in urination was divided into four degrees: none, mild, moderate, and severe, which was a subjective grading system for patients suffering from urinary retention as used in previous publications (22). The PGI-I assessment has potential scores of 1–7 representing much better, moderately better, a little better, no change, a little worse, moderately worse and much worse respectively, with a lower score corresponding to more symptom improvement (33). Symptomatic UTI was diagnosed by positive urine culture in a child with isolated fever or urinary symptoms suggestive of UTI. Positive urine culture was defined by the presence of >10⁵ colony forming units of a single organism per milliliter from a clean catch urine sample (34). Recurrent symptomatic UTI was defined as two or more episodes of symptomatic UTI within 24 weeks (35). The Rome IV criteria for children of a developmental age of at least 4 years were used to diagnose stool retention; stool retention was considered present when the child presented at least two of the six symptoms for a minimum of 1 month (36). Hydroureter and Hydronephrosis were scanned by abdominal ultrasound.

As different participants were likely to respond to different EA treatment duration, the response rate was also regarded as a time-to-event outcome in our study. Time-to-event analysis was used to explore the association between the response rate and EA treatment duration.

Considering that the recovery of bladder voiding function could be affected by the factors of baseline PVR and CUR duration of patients, subgroup analysis was performed to explore the association between the factors and the primary outcome or time-to-event outcome, and to figure out the potential factors affecting the efficiency of EA in treating CUR caused by LMNL following lumbosacral surgeries among pediatric patients (14, 37, 38).

All adverse events (AEs) were appropriately managed, monitored, and documented by investigators, and were categorized by the acupuncturist and urologists as either acupuncture related or not acupuncture related within 24 h of occurrence. Serious AEs were immediately reported to the principal investigator (L.Z.S.) and the institutional review board at the clinical sites within 24 h.

Skewed quantitative variables are expressed as median (IQR), and normally distributed variables as mean (SD). The distributions of quantitative variables were assessed for normality by use of the skewness and kurtosis test, and P ≥ 0.05 was thought normally distributed. Paired t-test or paired Fisher exact test were used to compare pre- and post-treatment data for quantitative and binary variables, respectively. Comparison of binary outcomes between subgroups was performed by independent-samples t-test or Fisher exact test. Time-to-event outcome was summarized by Kaplan–Meier estimates. Comparison of time-to-event outcome between subgroups was performed by Log-rank test. Two-sided P value less than 5% was considered significant. Statistical analyses were conducted using STATA software version 14.0 (Stata Corp, TX) for Windows.

From August 5, 2017 to July 31, 2022, a total of 21 patients were screened, of whom, 7 patients were excluded because of brain tumor resection surgeries (n = 1), neuritis (n = 2), myelitis (n = 2), treatment duration less than 1 week (n = 2) (Supplementary Figure S1). A total of 14 participants (mean [SD] age, 12 [4] years; 4 boys [29%]; 10 girls [71%]) were included in the analysis. The most common etiology was the tethered cord syndrome surgery in 6 patients (43%), followed by the lumbosacral vertebra fracture surgery in 5 patients (36%). Lesion location focused on cauda equina [12 participants (86%)] and sacral cord [2 participants (14%)]. Baseline characteristics of 14 participants were shown in Table 1.

After EA treatment, the response rate was 71% (10/14) at the 12th week (Table 2). Around 29% of participants (4/14) achieved a reduction of 90%–100% in PVR from baseline (Figure 1A). As shown in Figure 1B, 4 participants (29%) had responded to EA treatment within 4 weeks. Kaplan–Meier survival curve estimated that 50% participants responded to 8-week EA treatment and depicted a negative trend between the proportion of non-responders and EA treatment duration (Figure 2A).

As displayed in Table 2, at the 12th week, PVR significantly reduced by 246 (95% CI: 80, 412) ml, with an average of 68% reduction from baseline (P < 0.001). Likewise, the proportion of participants with severe difficulty in urination was 21% (3/14) with an average of 71% reduction from baseline (P < 0.001). Regarding the proportion of participants with stool retention, it just reduced by 14% from baseline (P = 0.625). After EA treatment, PGI-I score of 1 or 2 corresponding to much better or moderately better was seen in 9 participants (64%). The proportion of participants with catheter removal was 50% (7/14) at the 12th weeks, and none of these participants experienced re-catheterization in the 24-week follow-up. None of the 14 participants had evidence of recurrent UTI, newly emerging hydroureter or hydronephrosis.

No significant differences were detected in subgroups of baseline PVR and CUR duration in terms of the primary outcome (Supplementary Table S1). However, regarding time-to-event outcome, as illustrated by Figures 2B,C, participants with baseline PVR ≥300 ml and CUR duration ≥12 months experienced a longer EA treatment duration to attain the response rate of 50%, compared with participants whose PVR <300 ml and CUR duration <12 months (median value: 12 weeks vs. 8 weeks, 12 weeks vs. 4 weeks, respectively). It should be noted that although median EA treatment duration was differentiated between subgroups, P value was not significant attested by Log-rank test.

No episode of serious AEs was reported. Acupuncture-related AEs were infrequent, mild, and transient, which occurred in 2 participants (14%), including sharp pain in 1 participant and localized pigmentation in 1 participant of all cases. No participants discontinued EA treatment because of AEs.

This was the first clinical study to evaluate the efficacy and safety of EA in treating CUR caused by LMNL, following lumbosacral surgeries in pediatric patients. This study drew a conclusion that EA could reduce PVR and facilitate the recovery of bladder voiding function. But it has several limitations. Firstly, urodynamic examination was not conducted although PVR was regarded as the primary objective indicator of the outcome. Secondly, control group was not set as this study was to observe the effectiveness of EA in advance. Sham EA control should be set in further study to eliminate the possible interferences of placebo effects and natural course of disease. Thirdly, this was a single-center case series study of a small sample size, and a multi-center randomized controlled trial of a lager sample size should be designed in the future.