Low-intensity pulsed ultrasound for erectile dysfunction: a systematic review and limited quantitative synthesis

Introduction

Background

Erectile dysfunction (ED) is a common male sexual disorder with important psychosocial, relationship, quality-of-life, and cardiometabolic implications (1-10). Its burden increases with age and is closely associated with vascular risk factors, diabetes, obesity, hypertension, dyslipidemia, smoking, and cardiovascular disease (4-10). Standard management includes lifestyle and risk-factor optimization, phosphodiesterase type 5 inhibitors (PDE5i), vacuum erection devices, intracavernosal or intraurethral therapies, testosterone therapy in selected men, psychosexual interventions, and penile prosthesis implantation (11-15). These options are clinically useful.

Rationale and knowledge gap

However, unmet needs remain in men with contraindications to drug therapy, inadequate PDE5i response, poor adherence, major cardiometabolic comorbidity, diabetes-related or post-treatment ED, chronic prostatitis/chronic pelvic pain syndrome (CP/CPPS)-associated ED, or a preference for non-invasive restorative approaches. Validated patient-reported instruments, including the International Index of Erectile Function (IIEF), its erectile-function domain (IIEF-EF), the five-item IIEF (IIEF-5), and minimal clinically important difference (MCID) thresholds, provide standardized outcomes for ED trials (16-18). For device-based therapies, however, questionnaire-based improvement alone is insufficient to establish restorative biological efficacy and should be interpreted alongside objective vascular assessment, durability, and standardized safety surveillance. Low-intensity extracorporeal shockwave therapy (Li-ESWT) has been studied more extensively than low-intensity pulsed ultrasound (LIPUS), but the modalities are not interchangeable. Clinically, Li-ESWT delivers discrete shockwave impulses with an abrupt pressure rise, high peak pressure relative to LIPUS, short pulse duration, energy-flux-density reporting, and cavitation-related effects. LIPUS instead applies repeated low-intensity pulsed acoustic energy described by frequency, pulse repetition frequency, pulse width, duty cycle, intensity definition, and probe configuration. Thus, efficacy, safety, and dosing cannot be transferred between the modalities. Scroppo et al. reported improved penile hemodynamic parameters after Li-ESWT, supporting a possible neovascular response in vasculogenic ED (19); this clinically relevant Li-ESWT finding provides context but was not pooled with LIPUS evidence. These biophysical differences affect tissue interaction, mechanotransduction, dose reporting, and potential biological response (20-26). Methodological standards for systematic review reporting, appraisal, risk-of-bias assessment, certainty rating, narrative synthesis, and publication-bias interpretation have also been formalized (27-34). Clinical LIPUS protocols differ in device platform, acoustic output, pulse parameters, probe geometry, treatment site, weekly frequency, total sessions, concomitant therapy, and follow-up (35-40). This heterogeneity is the central reason why a focused LIPUS-specific synthesis is needed; indiscriminate pooling could confuse device efficacy with placebo-like response, schedule effects, adjunctive treatment, or patient selection.

Objective

This review aimed to evaluate the short-term efficacy and safety of LIPUS for ED compared with sham or other clinically relevant controls; examine limited protocol-related evidence, including fixed-session treatment frequency and combination/adjuvant therapy; and map protocol heterogeneity, source-level data completeness, risk of bias, and evidence certainty to inform standardized future trials. Li-ESWT-only studies were excluded to preserve intervention homogeneity and avoid conflating distinct energy-based modalities. We present this article in accordance with the PRISMA reporting checklist (27) (available at https://tau.amegroups.com/article/view/10.21037/tau-2026-0548/rc).

Methods

Study design and registration

This systematic review with limited quantitative synthesis and protocol-focused evidence mapping was conducted in accordance with AMSTAR 2 principles (28); the Cochrane Handbook guided synthesis methods (29,34). Ethical approval and informed consent were not applicable because only published aggregate data were analyzed. This systematic-review project was re-registered in the Open Science Framework (OSF) Registries on 2 September 2026 [updated OSF-ID: 7b9ed; (https://osf.io/7b9ed/) (https://osf.io/7b9ed/)]. The original registration record (former ID: 7s9kF), which was prospectively completed on 28 May 2026 prior to formal literature retrieval and synthesis, became inaccessible due to unexpected technical link failure; no core protocol items were altered during re-registration. Our formal batch literature search was completed on 8 June 2026 prior to data synthesis. The registered protocol specified the PICOS (participants, interventions, comparators, outcomes, and study designs) criteria, search strategies, extraction forms, risk-of-bias tools, Grading of Recommendations Assessment, Development and Evaluation (GRADE) framework, and statistical plan. No eligibility criteria, outcomes, or effect estimates were changed after data extraction; protocol-status details are reported in Table S1.

Eligibility criteria

Eligibility criteria were defined using the PICOS framework. Participants were adult men aged 18 years or older with ED diagnosed using validated questionnaires, physician assessment, or study-defined clinical criteria. Eligible etiologies included vasculogenic, mixed, metabolic, diabetes-related, age-related, post-treatment, PDE5i-refractory, and chronic prostatitis/chronic pelvic pain syndrome (CP/CPPS)-associated ED when erectile-function outcomes were extractable. Eligible interventions were LIPUS delivered as monotherapy or in combination with conservative therapy, including PDE5i, alpha-blockers, or biofeedback-assisted electrical stimulation. Comparators included sham therapy, alternative LIPUS schedules, drug therapy, LIPUS alone, or conservative non-surgical comparators. Li-ESWT-only studies were excluded a priori. This decision was made to preserve intervention homogeneity rather than to dismiss the relevance of Li-ESWT in ED research. LIPUS delivers pulsed ultrasound mechanical energy with device-specific frequency, pulse repetition frequency, pulse width, duty cycle, and acoustic output. In contrast, Li-ESWT uses shockwaves with a steep pressure rise, short pulse duration, different peak-pressure and energy-flux-density characteristics, and a distinct cavitation and mechanotransduction profile. These differences affect tissue interaction, dose reporting, treatment targeting, and potential biological response. Pooling Li-ESWT with LIPUS would therefore combine physically and mechanistically distinct interventions and could obscure modality-specific efficacy, safety, and protocol requirements. Primary outcomes were changes in erectile-function scores, mainly IIEF-EF or IIEF-5. Secondary outcomes included MCID or study-defined clinical response, Erection Hardness Score (EHS), Sexual Encounter Profile (SEP) outcomes, Global Assessment Question (GAQ), quality-of-life or relationship measures, penile color duplex Doppler ultrasonography (CDDU), adverse events (AEs), discontinuation, and durability of response. Eligible designs included randomized controlled trials, randomized schedule-comparison studies, prospective randomized clinical studies, and retrospective comparative clinical studies with extractable data. Animal studies, in vitro studies, case reports, reviews, editorials, Li-ESWT-only studies, non-ED populations, duplicate abstracts, and preliminary reports without distinct extractable clinical data were excluded.

Information sources and search strategy

A structured search was performed in PubMed/MEDLINE, Embase, Web of Science, Cochrane Library, Scopus, ClinicalTrials.gov, the World Health Organization International Clinical Trials Registry Platform (WHO ICTRP), China National Knowledge Infrastructure (CNKI), Wanfang Data, the VIP Journal Integration Platform (VIP), and the Chinese Biomedical Literature Database (SinoMed) from inception to 8 June 2026. English and Chinese records were screened without language restriction. Search concepts combined ED terms with LIPUS terms and device-related variants. Full database-specific syntax, fields, date limits, interfaces, export counts, and registry procedures are provided in Table S2.

All searches were completed in one batch on 8 June 2026 after the original prospective OSF registration and prior to the OSF re-registration. No publication-type, study-design, or language filters were applied at export. Records were imported into EndNote for deduplication; eligibility exclusions were made only during manual title/abstract and full-text screening. Chinese dissertations and conference records were retrieved and screened, but reports without complete extractable clinical data were excluded according to the prespecified criteria. Reference lists and citations of eligible studies and relevant reviews were screened independently by two reviewers. Database and trial-registry searches yielded 1,982 records, and reference/citation screening yielded 17, for 1,999 records before deduplication. Search logs, exports, deduplication records, screening logs, and extraction forms were retained. Trial-registry and bibliographic records were matched using identifiers, titles, interventions, sponsors, populations, and author/institution overlap. Detailed retrieval records and processing rules are reported in Table S2.

Study selection and data extraction

Two independent reviewers completed title/abstract screening and full-text eligibility assessment according to the finalized PICOS (participants, interventions, comparators, outcomes, and study designs) criteria. Retrieved citations were imported into EndNote for deduplication before screening. Disagreements were resolved by discussion, and unresolved discrepancies were adjudicated by a third senior reviewer. Inter-rater reliability was excellent for title/abstract screening (kappa =0.87) and full-text eligibility screening (kappa =0.91). When multiple reports described the same or substantially overlapping cohort, the most complete peer-reviewed full-text publication was retained. A 2018 Journal of Sexual Medicine supplement abstract was not counted separately because it was considered a probable preliminary or duplicate report of the Cui 2019 multicenter trial (35). Two reviewers independently extracted study identifier, country, design, population, group size, baseline erectile-function status, ED severity, age, ED duration, ED etiology, cardiometabolic comorbidities, diabetes status, hypertension, cardiovascular disease, smoking history, PDE5i use and responsiveness, baseline five-item International Index of Erectile Function (IIEF-5)/International Index of Erectile Function erectile-function domain (IIEF-EF), key eligibility restrictions, treatment protocol, comparator, follow-up, erectile-function outcomes, clinically meaningful response, penile Doppler parameters, AEs, discontinuation, and funding or manufacturer disclosures where available. Extracted variables were classified as directly reported, calculated from reported numbers, inferred from eligibility criteria, or not reported. Unavailable values were marked as not reported (NR) and were not imputed. For AEs, group-specific denominators and event counts were extracted only when reported or extractable from the original study; otherwise, the relevant cells were marked as NR. Adverse-event causality was recorded only as assessed in the original studies, and this review did not independently adjudicate AE causality. A revised source-location matrix was added as Table S3 to identify whether each data element came from baseline tables, methods, results, trial design statements, eligibility criteria, or calculations from reported denominators.

Risk of bias and certainty of evidence

Risk of bias was assessed according to study design. Randomized studies were evaluated using the Cochrane RoB 2 framework (30), whereas non-randomized comparative studies were evaluated using ROBINS-I (31). For RoB 2, judgments considered randomization, allocation concealment, deviations from intended interventions, missing outcome data, outcome measurement, and selection of the reported result. For ROBINS-I, judgments considered confounding, participant selection, intervention classification, deviations from intended intervention, missing data, outcome measurement, and selective reporting. Particular attention was given to baseline ED severity, PDE5i responsiveness, concomitant tadalafil or non-pharmacological therapy, comorbidity burden, blinding, subjective endpoints, and treatment-selection mechanisms. Tool-specific terminology was retained: RoB 2 was reported as low risk, some concerns, or high risk; ROBINS-I was reported as low, moderate, serious, critical, or no information. Detailed domain-level judgments are presented in Table S4, with a concise handling summary in Table 1. GRADE was used to rate certainty of evidence for key outcomes (32). Certainty was downgraded by one or two levels, as appropriate, for risk of bias, inconsistency, indirectness, imprecision, and publication bias. Single-trial evidence, modest information size, subjective endpoints, short follow-up, non-prospective or retrospective allocation, residual confounding, clinically incompatible adjunct strategies, incomplete objective assessment, and geographic restriction to Chinese study populations were explicitly considered. The reason and level of downgrading for each outcome are summarized in Table 2 and Table S5.

Statistical analysis

Continuous outcomes were summarized as mean differences (MDs) with 95% confidence intervals (CIs). Dichotomous outcomes were summarized as risk ratios (RRs) with 95% CIs. Single-study comparisons were interpreted as unpooled estimates, and I2 was not calculated because between-study heterogeneity is not applicable to one study. Any software-generated graphical marker in a single-study display should be read only as the individual study estimate, not as a pooled meta-analytic effect. For non-inferiority trials, we extracted the source article prespecified non-inferiority margins, primary analysis population, sensitivity analysis population, and whether reported confidence limits satisfied those margins. Non-inferiority was not reinterpreted as equivalence. Chen 2022 was handled as a source-defined randomized non-inferiority schedule-comparison trial. The source trial prespecified two margins: an RR margin of 0.80 for treatment-response rate and a −2.5-point margin for between-group IIEF-EF difference. The primary analysis used the intention-to-treat (ITT) population, and the per-protocol (PP) sensitivity analysis was reported as consistent. The review did not independently validate the clinical acceptability of these margins or perform a de novo non-inferiority test; it therefore adopted the source conclusion only for the fixed-total-session frequency comparison and did not infer equivalence, superiority, biological interchangeability, or optimal dosing. Clinical heterogeneity was prespecified across device platform, manufacturer, acoustic intensity definition, pulse repetition frequency, pulse width, duty cycle, treatment site, weekly frequency, total sessions, ED severity, PDE5i responsiveness, cardiometabolic risk profile, concomitant therapy, comparator type, and follow-up interval. I2 was interpreted only when at least two clinically comparable studies contributed to the same comparison. I2 values below 25% were considered low, 25–50% moderate, and above 50% substantial heterogeneity. A fixed-effect model was planned for low-heterogeneity comparisons, whereas a random-effects model was planned when I2 exceeded 50% or when clinically meaningful heterogeneity was expected. Comparisons with substantial statistical or clinical heterogeneity were treated as exploratory rather than confirmatory. Quantitative pooling was not performed across clinically different questions, including sham-controlled efficacy, active treatment-frequency comparison, and combination/adjuvant therapy. When comparators, populations, adjunctive treatments, outcome definitions, or follow-up windows were incompatible, findings were synthesized narratively using principles consistent with synthesis without meta-analysis (33). Exploratory graphical displays of clinically heterogeneous adjunctive studies were retained only to show the direction and uncertainty of individual estimates; no pooled adjunctive effect was used for clinical inference. Funnel plots and Egger-type tests were not performed because fewer than 10 studies contributed to each outcome comparison (34). Subgroup analyses were considered for ED severity, etiology, diabetes status, PDE5i responsiveness, metabolic or vascular comorbidity profile, age, CP/CPPS status, device platform, treatment site, acoustic intensity, and weekly frequency. However, formal pooled subgroup estimates and interaction tests were not performed because strata were frequently single-study based, involved incompatible comparators, or differed substantially in concomitant therapy and acoustic protocol. All subgroup information was therefore used only for descriptive study-level mapping and hypothesis generation, not for clinical efficacy inference or patient-selection recommendations.

Artificial intelligence (AI)-assisted tool use

OpenAI ChatGPT (GPT-5.6 Thinking; OpenAI; accessed 24 July 2026) was used only to improve English-language expression and to check grammar, spelling, punctuation, and sentence structure.

Results

Study selection and study characteristics

The study-selection process is summarized in Figure 1. The updated search retrieved 1,982 records from databases and trial registries, and 17 additional records were identified through reference or citation screening, yielding 1,999 records before deduplication. After removal of 624 duplicates, 1,375 records underwent title/abstract screening, 1,286 were excluded, and 89 full-text articles were assessed. Eighty-three full-text articles were excluded for four protocol-defined reason categories: Li-ESWT-only or non-LIPUS interventions (n=39), animal/in vitro studies or non-human ED research (n=21), reviews, case reports, conference abstracts, or reports without extractable original data (n=16), and duplicate or overlapping reports (n=7). Six clinical studies involving 659 participants met eligibility criteria (35-40). The characteristics and synthesis role of each included study are presented in Table 3. The evidence architecture is summarized in Figure 2, and the clinical and protocol landscape is summarized in Figure 3. All were conducted in China, which is the primary external-validity limitation and restricts direct applicability to Chinese study populations. Four studies used randomized designs and two were retrospective comparative studies. The included populations ranged from mild-to-moderate ED to severe ED and ED associated with CP/CPPS.

Figure 1 PRISMA flow diagram. Searches of databases and trial registries yielded 1,982 records, and manual reference/citation screening identified 17 additional records. After removal of 624 duplicates, 1,375 records were screened, 89 full-text articles were assessed, and six clinical studies were included. Five studies contributed extractable comparison-specific quantitative estimates. CNKI, China National Knowledge Infrastructure; ED, erectile dysfunction; Li-ESWT, low-intensity extracorporeal shockwave therapy; LIPUS, low-intensity pulsed ultrasound; PRISMA, Preferred Reporting Items for Systematic Reviews and Meta-Analyses.

Figure 2 Study selection and evidence architecture. The figure separates sham-controlled efficacy, treatment-frequency evidence, and exploratory combination/adjuvant evidence. Clinically heterogeneous adjunctive evidence was synthesized narratively in the main text. IIEF-5, five-item International Index of Erectile Function; LIPUS, low-intensity pulsed ultrasound; MCID, minimal clinically important difference; MD, mean difference; PICOS, participants, interventions, comparators, outcomes, and study designs; PRISMA, Preferred Reporting Items for Systematic Reviews and Meta-Analyses; RR, risk ratio.

Figure 3 Clinical and protocol landscape of included LIPUS-ED studies. Filled circles indicate extractable data; open circles indicate data that were not extractable or not reported. CDDU, color duplex Doppler ultrasonography; CP/CPPS, chronic prostatitis/chronic pelvic pain syndrome; ED, erectile dysfunction; IIEF, International Index of Erectile Function; LIPUS, low-intensity pulsed ultrasound; NR, not reported; RCT, randomized controlled trial.

Baseline clinical characteristics were extracted into a structured study-level matrix (Table 4). The revised table reports age, ED duration, ED severity, diabetes, hypertension, cardiovascular disease, smoking history, PDE5i use and responsiveness, baseline IIEF-5/IIEF-EF, key eligibility restrictions, and descriptive subgroup dimensions supported by each study. Where source reports did not directly provide a variable, the value was marked as NR or explicitly treated as inferred from eligibility criteria rather than imputed. These variables help explain between-study differences, but they do not justify pooled subgroup interaction testing because strata were often represented by a single study or involved incompatible comparators and treatment protocols.

LIPUS treatment protocols

Treatment protocols varied substantially across device platform, manufacturer, acoustic output definition, acoustic intensity, pulse repetition frequency, pulse width, duty cycle, treatment site, weekly frequency, total sessions, and concomitant therapy. Table 5 was therefore revised to standardize all reported acoustic intensity values to mW/cm2 using the nominal arithmetic conversion 1 W/cm2 =1,000 mW/cm2. After conversion, the reported active intensities ranged from 200–300 mW/cm2 in WBL-ED studies, whereas the Lanyue dual-probe study reported 2,500 and 1,250 mW/cm2 and the custom-device study reported 3,000 mW/cm2. These converted values improve readability but should not be interpreted as equivalent biological doses, because intensity definition, measurement basis, probe geometry, pulse structure, output calibration, coupling, and treatment field were not harmonized. Four of six studies used the WBL-ED platform, but with non-identical settings and schedules; Li 2024 and Yang 2024 used distinct platforms. Consequently, device and parameter heterogeneity is a likely contributor to clinical heterogeneity and precludes inference about an optimal dose, anatomical delivery pattern, cumulative exposure, or standardized regimen.

Response definitions and evidence comparability

The response definitions were not fully interchangeable across studies (Table 6). The main efficacy estimate therefore relied on the single sham-controlled trial, whereas schedule-comparison and adjunctive studies were interpreted within their own clinical and methodological contexts. Response rates based on different scales, thresholds, follow-up windows, or adjunctive regimens were not treated as exchangeable clinical endpoints.

Subgroup and baseline-stratified analysis feasibility

The structured baseline dataset allowed descriptive mapping across ED severity, PDE5i responsiveness, metabolic/vascular comorbidity profile, age, and CP/CPPS status. Cui 2019, Xia 2020, Chen 2022, and Yang 2024 mainly represented mild-to-moderate ED; Li 2024 represented ED with CP/CPPS and a higher PDE5i low-response burden; and Gao 2024 represented severe, predominantly PDE5i-nonresponsive ED with the highest metabolic and vascular risk burden. These categories improve clinical interpretation of between-study differences. However, they do not provide a valid basis for reliable subgroup effect modification or interaction testing. All subgroup findings in this review should be read as descriptive and hypothesis-generating rather than confirmatory; they should not be used for patient-selection recommendations.

Risk of bias

A protocol-oriented evidence map covering the three distinct evidence domains is provided in Table 7. Risk-of-bias judgments are summarized in Table 1 and Figure 4, with expanded domain-level details in Table S4. The Cui 2019 sham-controlled randomized trial was judged as having some concerns, mainly because direct efficacy evidence came from a single trial with attrition and limited allocation-concealment reporting (35). Xia 2020 and Chen 2022 were judged as having some concerns because they were active schedule-comparison studies with subjective endpoints and no sham control; Chen 2022 was also open-label, which may influence patient-reported outcomes (36,37). Li 2024 was judged as high risk because of its small sample, open pragmatic three-arm design, short follow-up, and combination-treatment context (38). Gao 2024 and Yang 2024 were judged as serious risk under ROBINS-I because retrospective treatment allocation introduced residual confounding by baseline ED severity, comorbidity burden, PDE5i responsiveness, concomitant therapy, and treatment-selection mechanisms (39,40).

Figure 4 Tool-specific risk-of-bias summaries. (A) RoB 2 domain-level judgments for the four randomized studies. (B) ROBINS-I domain-level judgments for the two retrospective comparative studies. The harmonized visual legend uses Low, Some concerns, High, and Serious for readability; these categories should not be interpreted as numerically equivalent across RoB 2 and ROBINS-I. Domain labels are defined within each panel, and italic notes indicate the principal bias sources contributing to domain-level judgments. Expanded justifications are provided in Table S4. CDDU, color duplex Doppler ultrasonography; ROBINS-I, Risk Of Bias In Non-randomized Studies - of Interventions; RoB 2, revised Cochrane risk-of-bias tool for randomized trials.

Sham-controlled efficacy

The principal direct efficacy evidence came from the sham-controlled Cui 2019 trial (35). At 12 weeks, LIPUS was associated with greater short-term improvement in patient-reported IIEF-5 score and a higher clinically meaningful response rate than sham treatment (Figures 5,6). Because this comparison was based on one sham-controlled trial, it was interpreted as a single-study estimate; no confirmatory pooled overall effect or I2 statistic was calculated. These findings support a possible short-term subjective benefit in selected Chinese men with mild-to-moderate ED, but certainty is limited by single-trial evidence, attrition, short follow-up, modest information size, and the absence of adequate objective corroboration. The main sham-controlled efficacy results are summarized in Figure 5.

Figure 5 Single-study estimate of erectile-function score change. The estimate is based on the single sham-controlled Cui 2019 trial and should not be interpreted as a pooled meta-analysis. CI, confidence interval; IIEF-5, five-item International Index of Erectile Function; IIEF-EF, International Index of Erectile Function erectile-function domain; LIPUS, low-intensity pulsed ultrasound.

Figure 6 Forest plot of clinically meaningful response rate. Panel A shows LIPUS versus sham treatment, Panel B shows the Chen 2022 treatment-frequency estimate within the source non-inferiority framework, and Panel C displays individual clinically heterogeneous combination/adjuvant estimates for transparency; no pooled estimate was calculated. MCID or study-defined improvement criteria were used according to the source study. CI, confidence interval; LIPUS, low-intensity pulsed ultrasound; MCID, minimal clinically important difference.

Treatment-frequency evidence EHS and other patient-reported sexual-function outcomes were reported across most included studies, but instruments, comparators, and follow-up intervals differed (35-40). Two randomized studies involving 363 participants compared three sessions weekly with two sessions weekly when the total number of sessions was fixed (36,37). This evidence addresses only treatment interval under a fixed-session framework. It does not identify the optimal acoustic dose, anatomical delivery pattern, total treatment course, cumulative energy exposure, device platform, or clinical setting in which one schedule should be preferred. Chen 2022 requires explicit non-inferiority interpretation. The source trial prespecified two non-inferiority margins: a treatment-response RR margin of 0.80 and a between-group IIEF-EF margin of −2.5 points. The treatment-response estimate was RR 0.99 (95% CI: 0.84 to 1.18), so the lower CI limit exceeded the RR margin (0.84>0.80). The reported lower confidence limit for the IIEFEF difference also exceeded the functional-score margin (−1.2>−2.5). Thus, Chen 2022 met its source-defined non-inferiority criteria for both key endpoints. Clinically, however, these margins require cautious interpretation. Published MCID estimates for the IIEF-EF domain are severity-dependent, and a −2.5-point margin may represent a potentially meaningful loss for some men with mild ED while being smaller than commonly reported MCID thresholds for more severe baseline dysfunction (18). The RR margin of 0.80 allows as much as a 20% relative reduction in treatment response, and the source article did not provide an independently verifiable ED-specific external justification for this threshold. Accordingly, this review reports the source-defined non-inferiority conclusion without independently validating the clinical acceptability of either margin. The open-label schedule comparison may also be vulnerable to expectation effects because the main outcomes were patient-reported. This conclusion applies only to the fixed-total-session comparison in the source trial, does not prove equivalence or superiority, and should not be generalized to other devices, populations, total courses, acoustic protocols, or biological effects. Combination/adjuvant therapy Exploratory studies of combination or adjuvant therapy should not be used for clinical decision-making (38-40). Across Li 2024, Gao 2024, and Yang 2024, adjunct strategies, ED severity, CP/CPPS status, concomitant drugs, study design, and follow-up intervals differed substantially. Therefore, the combination/adjuvant estimates shown in Figure 6 and Figure S1 are exploratory visual summaries for transparency only and should not be interpreted as a confirmatory pooled effect. These data do not provide reliable evidence that LIPUS combined with PDE5i, alpha-blocker therapy, or biofeedback electrical stimulation is superior to LIPUS monotherapy. Penile hemodynamic outcomes Penile hemodynamic outcomes were reported in the Gao 2024 retrospective cohort, but complete CDDU data were available only in a subset of six participants (39). These data were not collected as a standardized, prospectively powered objective endpoint and were not corroborated across trials. No included randomized sham-controlled trial demonstrated improvement in standardized penile Doppler parameters. Given the extremely small subset, retrospective design, and absence of uniform vascular assessment, the CDDU findings are hypothesis-generating only. They should not be used as clinical evidence that LIPUS definitively improves penile arterial inflow, venous occlusion, vascular remodeling, or tissue restoration.

Safety assessment and AEs

AEs were documented across all six included studies (35-40), but AE definitions, severity grading, causality methods, participant denominators, and surveillance schedules were not standardized. Reported reactions were predominantly mild and transient, including penile numbness, warmth, distension pain, erythema, pruritus, headache, dizziness, low back soreness, insomnia, and urinary frequency (Table 8). During short-term follow-up (no longer than 12 weeks), no serious device-related AE and no AE-related discontinuation were reported. However, the lack of standardized AE collection and causality adjudication in the primary studies prevented definitive safety conclusions; this limitation reflects the available evidence rather than the review analysis. Arm-level denominator and ascertainment details are provided in Table S6. Current evidence cannot exclude rare, delayed, tissue-level, parameter-dependent, or device-specific harms.

Discussion

Key findings

This systematic review with limited quantitative synthesis provides a focused evaluation of the available clinical evidence for LIPUS in adult men with ED. The phrase limited quantitative synthesis is used deliberately because the principal direct efficacy evidence came from one multicenter double-blind sham-controlled randomized trial, and most other studies addressed different clinical questions (35-40). The available evidence suggests possible short-term patient-reported benefit beyond a sham control in selected men with mild-to-moderate ED. However, this inference is fragile because the core sham-controlled evidence includes only 120 participants, all clinical studies were conducted in China, most endpoints were subjective, objective hemodynamic data were almost absent, and follow-up did not extend beyond 12 weeks. The present conclusions should therefore be considered applicable only to Chinese study populations at this stage, and cross-ethnic extrapolation to non-Chinese or multiethnic populations requires caution. A central methodological implication is that the evidence base addresses distinct clinical questions rather than one homogeneous treatment comparison. Sham-controlled efficacy, active treatment-frequency comparison, and adjunctive combination strategies differ in purpose, design, comparator, susceptibility to bias, and clinical interpretability. These domains should not be pooled indiscriminately. When total sessions were fixed, Chen et al. [2022] met its source-defined non-inferiority margins for three sessions weekly versus two sessions weekly, and Xia et al. [2020] showed no material score disadvantage for the more frequent schedule (36,37). However, the clinical acceptability of the RR 0.80 margin and the -2.5-point IIEF-EF margin was not independently validated by this review. Because IIEF-EF MCID values vary according to baseline ED severity (18), and because open-label subjective outcomes can bias results toward apparent similarity, the Chen et al. [2022] finding should be interpreted as schedule non-inferiority only within the original source-defined framework. It may be relevant for calendar-course planning, but it does not prove equivalence, superiority, optimal acoustic dose, optimal pulse mode, optimal treatment site, optimal total exposure, or a generalizable dosing strategy. The evidence for LIPUS combined with PDE5i, alpha-blocker therapy, or biofeedback-assisted electrical stimulation remains too heterogeneous for confirmatory pooling (38-40). The adjunctive studies differed in ED severity, CP/CPPS status, PDE5i responsiveness, concomitant therapy, study design, and follow-up. As a result, the main synthesis treats these data as directionally inconsistent and clinically heterogeneous evidence rather than as proof that combined or adjunctive therapy is superior to LIPUS alone. This conservative interpretation is important because a pooled estimate across biologically and clinically different adjuncts could create a misleading impression of a generalizable treatment effect.

Strengths and limitations

This review has several strengths: prospective OSF registration, explicit PRISMA accounting, bilingual searching, separation of clinically distinct comparisons, design-appropriate use of RoB 2 and ROBINS-I, GRADE assessment, source-location checks, and avoidance of inappropriate pooling or funnel-plot testing. Nevertheless, the evidence has major limitations. First, all six clinical studies were conducted in China, restricting external validity across ethnic, cultural, healthcare, and cardiometabolic contexts. Second, the principal sham-controlled trial included only 120 participants, and complete objective hemodynamic data were available in only six participants from one retrospective cohort; sham-controlled efficacy was therefore treated as a single-study estimate and Doppler findings as hypothesis-generating. Third, protocols varied in device platform, acoustic intensity and its definition, pulse parameters, probe configuration, treatment site, session schedule, cumulative exposure, and concomitant therapy. Converting reported intensity units to mW/cm2 did not make nominal settings biologically interchangeable. Fourth, follow-up did not establish medium- or long-term durability or delayed safety. Fifth, adverse-event definitions, denominators, event counts, causality methods, and surveillance schedules were inconsistently reported, and this review did not independently adjudicate causality. These limitations were addressed by downgrading certainty, separating evidence domains, and limiting clinical claims.

Implications for future trials

Future studies should be international, multicenter, adequately powered, double-blind, sham-controlled randomized trials across clinically important ED etiologies and severity groups. Baseline PDE5i responsiveness, diabetes, cardiovascular risk, ED duration, and comorbidity should be prespecified for stratification or adjustment. Protocols and analysis plans should be registered before enrollment. Intervention descriptions should follow the Template for Intervention Description and Replication (TIDieR) framework (41) and report device/manufacturer, frequency, pulse repetition frequency, pulse width, duty cycle, original and standardized intensity units, intensity/output definition, probe geometry, calibration, treatment sites, session duration, weekly frequency, total sessions, and cumulative exposure. Subjective IIEF outcomes should be paired with standardized objective CDDU endpoints, prespecified MCID thresholds, validated AE collection and causality procedures, and follow-up at 3, 6, and 12 months.

Comparison with similar research

Clinical positioning remains unresolved. PDE5i therapy remains the established first-line pharmacological option for many men with ED, with broader evidence and clearer clinical pathways. Vacuum devices, injection therapy, psychosexual interventions, hormonal therapy in selected men, and penile prosthesis implantation also have defined roles in guideline-based care. LIPUS is best regarded as an investigational restorative or adjunctive device-based therapy rather than a replacement for established treatments. Li-ESWT has a larger clinical literature than LIPUS, but its effects, protocol assumptions, and safety profile cannot be directly transferred to LIPUS because the two modalities are physically and mechanistically distinct (20-26).

Explanations of findings

The biological rationale for LIPUS is plausible but remains clinically unconfirmed. Erectile function depends on endothelial relaxation, cavernosal smooth-muscle integrity, arterial inflow, veno-occlusive function, and neural signaling. Preclinical studies suggest that LIPUS may influence endothelial repair, angiogenesis, nitric oxide-cGMP signaling, fibrosis, inflammation, and neural repair (42-45). A preclinical and hypothetical mechanistic schematic is shown in Figure 8. However, subjective questionnaire-based improvement in a single sham-controlled trial cannot demonstrate human penile vascular remodeling, neural restoration, or durable tissue-level repair. No sham-controlled randomized trial included in this review demonstrated objective improvement in standardized penile Doppler ultrasound parameters. In the current evidence base, complete objective Doppler data were available only in six participants from one retrospective cohort. Mechanistic and hemodynamic claims should therefore remain hypothesis-generating until prospectively powered trials include standardized CDDU or other objective vascular endpoints. Recent narrative reviews, protocol reports, and emerging-therapy reviews similarly support protocol standardization and further confirmatory trials (46-49). LIPUS should also remain clearly separated from Li-ESWT. Li-ESWT is usually conceptualized around shockwave-related mechanical stress, abrupt pressure changes, cavitation-related tissue effects, and downstream repair responses. LIPUS delivers pulsed ultrasound mechanical energy with different waveform, pulse structure, acoustic output, focal characteristics, and reporting requirements. Because these modalities are physically and mechanistically non-equivalent, Li-ESWT-only studies were excluded from this review and were not pooled with LIPUS studies.

Figure 8 Preclinical and hypothetical mechanistic schematic. The pathway summarizes preclinical plausibility but should not be interpreted as clinical proof of definitive hemodynamic improvement. LIPUS, low-intensity pulsed ultrasound; NO-cGMP, nitric oxide-cyclic guanosine monophosphate; sGC, soluble guanylate cyclase.

Implications and actions needed

For clinicians, the current evidence supports cautious interpretation rather than routine adoption. LIPUS should preferably be evaluated within well-designed clinical trials until durable efficacy, objective vascular effects, optimal protocol, delayed safety, and generalizability beyond Chinese study populations are established. Outside a trial setting, use should be regarded as investigational and should involve careful counseling regarding uncertainty, standardized protocol documentation, objective baseline assessment where feasible, active adverse-event monitoring, and continued use of guideline-based ED management. The available evidence does not justify replacing established ED therapies and does not establish that combination with PDE5i, alpha-blockers, or biofeedback electrical stimulation is superior to LIPUS monotherapy.

Conclusions

Based on current low-certainty evidence, LIPUS remains investigational and cannot be recommended for routine clinical use in ED at this time. One sham-controlled randomized trial suggests that it may improve short-term patient-reported erectile function and clinically meaningful response in selected Chinese men with mild-to-moderate ED. Evidence is insufficient for severe ED, diabetes-related ED, or combination/adjuvant regimens, and no sham-controlled randomized trial demonstrated objective hemodynamic benefit. With a fixed total number of sessions, Chen et al. [2022] met source-defined non-inferiority margins for three versus two weekly sessions; this did not establish equivalence, superiority, optimal dosing, or generalizability across devices, populations, or courses. All included studies were conducted in China, objective Doppler evidence was limited to a six-participant retrospective subset, and follow-up was no longer than 12 weeks. No serious device-related AE or AE-related discontinuation was reported, but AE collection and causality assessment were not standardized. International multicenter sham-controlled RCTs should use standardized acoustic reporting, objective hemodynamic endpoints, prespecified subgroups, transparent AE assessment, and 3-, 6-, and 12-month follow-up.

Figure 7 Publication bias not formally assessable. Funnel-plot asymmetry testing and Egger-type tests were not performed because fewer than 10 studies contributed to each outcome comparison.

Acknowledgments

During revision, OpenAI ChatGPT (GPT-5.6 Thinking; OpenAI; accessed 24 July 2026) was used only for English-language polishing and grammar checking.

Footnote

Reporting Checklist: The authors have completed the PRISMA reporting checklist. Available at https://tau.amegroups.com/article/view/10.21037/tau-2026-0548/rc

Peer Review File: Available at https://tau.amegroups.com/article/view/10.21037/tau-2026-0548/prf

Funding: None.

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://tau.amegroups.com/article/view/10.21037/tau-2026-0548/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. NIH Consensus Conference. Impotence. NIH Consensus Development Panel on Impotence. JAMA 1993;270:83-90. [PubMed]
2. Yafi FA, Jenkins L, Albersen M, et al. Erectile dysfunction. Nat Rev Dis Primers 2016;2:16003. [Crossref] [PubMed]
3. Shamloul R, Ghanem H. Erectile dysfunction. Lancet 2013;381:153-65. [Crossref] [PubMed]
4. Feldman HA, Goldstein I, Hatzichristou DG, et al. Impotence and its medical and psychosocial correlates: results of the Massachusetts Male Aging Study. J Urol 1994;151:54-61. [Crossref] [PubMed]
5. Ayta IA, McKinlay JB, Krane RJ. The likely worldwide increase in erectile dysfunction between 1995 and 2025 and some possible policy consequences. BJU Int 1999;84:50-6. [Crossref] [PubMed]
6. Prins J, Blanker MH, Bohnen AM, et al. Prevalence of erectile dysfunction: a systematic review of population-based studies. Int J Impot Res 2002;14:422-32. [Crossref] [PubMed]
7. Selvin E, Burnett AL, Platz EA. Prevalence and risk factors for erectile dysfunction in the US. Am J Med 2007;120:151-7. [Crossref] [PubMed]
8. Gandaglia G, Briganti A, Jackson G, et al. A systematic review of the association between erectile dysfunction and cardiovascular disease. Eur Urol 2014;65:968-78. [Crossref] [PubMed]
9. Dong JY, Zhang YH, Qin LQ. Erectile dysfunction and risk of cardiovascular disease: meta-analysis of prospective cohort studies. J Am Coll Cardiol 2011;58:1378-85. [Crossref] [PubMed]
10. Nehra A, Jackson G, Miner M, et al. The Princeton III Consensus recommendations for the management of erectile dysfunction and cardiovascular disease. Mayo Clin Proc 2012;87:766-78. [Crossref] [PubMed]
11. Burnett AL, Nehra A, Breau RH, et al. Erectile Dysfunction: AUA Guideline. J Urol 2018;200:633-41. [Crossref] [PubMed]
12. Salonia A, Capogrosso P, Boeri L, et al. European Association of Urology Guidelines on Male Sexual and Reproductive Health: 2025 Update on Male Hypogonadism, Erectile Dysfunction, Premature Ejaculation, and Peyronie's Disease. Eur Urol 2025;88:76-102. [Crossref] [PubMed]
13. Goldstein I, Lue TF, Padma-Nathan H, et al. Oral sildenafil in the treatment of erectile dysfunction. Sildenafil Study Group. N Engl J Med 1998;338:1397-404. [Crossref] [PubMed]
14. Hatzimouratidis K, Amar E, Eardley I, et al. Guidelines on male sexual dysfunction: erectile dysfunction and premature ejaculation. Eur Urol 2010;57:804-14. [Crossref] [PubMed]
15. Porst H, Burnett A, Brock G, et al. SOP conservative (medical and mechanical) treatment of erectile dysfunction. J Sex Med 2013;10:130-71. [Crossref] [PubMed]
16. Rosen RC, Riley A, Wagner G, et al. The international index of erectile function (IIEF): a multidimensional scale for assessment of erectile dysfunction. Urology 1997;49:822-30. [Crossref] [PubMed]
17. Cappelleri JC, Rosen RC, Smith MD, et al. Diagnostic evaluation of the erectile function domain of the International Index of Erectile Function. Urology 1999;54:346-51. [Crossref] [PubMed]
18. Rosen RC, Allen KR, Ni X, et al. Minimal clinically important differences in the erectile function domain of the International Index of Erectile Function scale. Eur Urol 2011;60:1010-6. [Crossref] [PubMed]
19. Scroppo FI, Pezzoni F, Gaeta F, et al. Li-Eswt improves hemodynamic parameters thus suggesting neoangiogenesis in patients with vascular erectile dysfunction. Int J Impot Res 2022;34:237-42. [Crossref] [PubMed]
20. Chung E, Lee J, Liu CC, et al. Clinical Practice Guideline Recommendation on the Use of Low Intensity Extracorporeal Shock Wave Therapy and Low Intensity Pulsed Ultrasound Shock Wave Therapy to Treat Erectile Dysfunction: The Asia-Pacific Society for Sexual Medicine Position Statement. World J Mens Health 2021;39:1-8. [Crossref] [PubMed]
21. Ergun O, Kim K, Kim MH, et al. Low-intensity shockwave therapy for erectile dysfunction. Cochrane Database Syst Rev 2025;7:CD013166. [Crossref] [PubMed]
22. Lu Z, Lin G, Reed-Maldonado A, et al. Low-intensity Extracorporeal Shock Wave Treatment Improves Erectile Function: A Systematic Review and Meta-analysis. Eur Urol 2017;71:223-33. [Crossref] [PubMed]
23. Yao H, Wang X, Liu H, et al. Systematic Review and Meta-Analysis of 16 Randomized Controlled Trials of Clinical Outcomes of Low-Intensity Extracorporeal Shock Wave Therapy in Treating Erectile Dysfunction. Am J Mens Health 2022;16:15579883221087532. [Crossref] [PubMed]
24. Vardi Y, Appel B, Kilchevsky A, et al. Does low intensity extracorporeal shock wave therapy have a physiological effect on erectile function? Short-term results of a randomized, double-blind, sham controlled study. J Urol 2012;187:1769-75. [Crossref] [PubMed]
25. Kitrey ND, Gruenwald I, Appel B, et al. Penile Low Intensity Shock Wave Treatment is Able to Shift PDE5i Nonresponders to Responders: A Double-Blind, Sham Controlled Study. J Urol 2016;195:1550-5. [Crossref] [PubMed]
26. Hayon S, Panken EJ, Bennett NE. Variations in Low Intensity Shockwave Treatment Protocols for Erectile Dysfunction: A Review of the Literature and Guide to Offering Treatment. World J Mens Health 2024;42:283-9. [Crossref] [PubMed]
27. Page MJ, McKenzie JE, Bossuyt PM, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ 2021;372:n71. [Crossref] [PubMed]
28. Shea BJ, Reeves BC, Wells G, et al. AMSTAR 2: a critical appraisal tool for systematic reviews that include randomised or non-randomised studies of healthcare interventions, or both. BMJ 2017;358:j4008. [Crossref] [PubMed]
29. Higgins JPT, Thomas J, Chandler J, et al. editors. Cochrane Handbook for Systematic Reviews of Interventions. Version 6.4. Cochrane; 2023.
30. Sterne JAC, Savovic J, Page MJ, et al. RoB 2: a revised Cochrane risk-of-bias tool for randomized trials. Cochrane; 2019.
31. Sterne JA, Hernán MA, Reeves BC, et al. ROBINS-I: a tool for assessing risk of bias in non-randomised studies of interventions. BMJ 2016;355:i4919. [Crossref] [PubMed]
32. Guyatt G, Oxman AD, Akl EA, et al. GRADE guidelines: 1. Introduction-GRADE evidence profiles and summary of findings tables. J Clin Epidemiol 2011;64:383-94. [Crossref] [PubMed]
33. Campbell M, McKenzie JE, Sowden A, et al. Synthesis without meta-analysis (SWiM) in systematic reviews: reporting guideline. BMJ 2020;368:l6890.
34. Sterne JA, Sutton AJ, Ioannidis JP, et al. Recommendations for examining and interpreting funnel plot asymmetry in meta-analyses of randomised controlled trials. BMJ 2011;343:d4002. [Crossref] [PubMed]
35. Cui W, Li H, Guan R, et al. Efficacy and safety of novel low-intensity pulsed ultrasound (LIPUS) in treating mild to moderate erectile dysfunction: a multicenter, randomized, double-blind, sham-controlled clinical study. Transl Androl Urol 2019;8:307-19. [Crossref] [PubMed]
36. Xia SJ, Chen HR, Li Z, et al. Efficacy and safety of low-intensity pulsed ultrasound at different intervals by mechanical force in treating erectile dysfunction: a preliminary study. Zhonghua Yi Xue Za Zhi 2020;100:1432-6. [Article in Chinese]. [Crossref] [PubMed]
37. Chen H, Li Z, Li X, et al. The Efficacy and Safety of Thrice vs Twice per Week Low-Intensity Pulsed Ultrasound Therapy for Erectile Dysfunction: A Randomized Clinical Trial. J Sex Med 2022;19:1536-45. [Crossref] [PubMed]
38. Li Z, Li D, Zu X, et al. Evaluation of efficacy and safety of low-intensity pulsed ultrasound in patients with concurrent erectile dysfunction and chronic prostatitis/chronic pelvic pain syndrome: a prospective, randomized controlled study. BMC Urol 2024;24:232. [Crossref] [PubMed]
39. Gao QQ, Wang J, Li DS, et al. Efficacy and safety of low-intensity pulsed ultrasound (LIPUS) combined with tadalafil in the treatment of severe erectile dysfunction: a retrospective cohort study. Transl Androl Urol 2024;13:2045-54. [Crossref] [PubMed]
40. Yang L, Yushanjiang S, Zhang S, et al. The efficiency of biofeedback electric stimulation therapy combined with low-intensity pulsed ultrasound in treating erectile dysfunction: a clinical study. Rev Int Androl 2024;22:17-24. [Crossref] [PubMed]
41. Hoffmann TC, Glasziou PP, Boutron I, et al. Better reporting of interventions: template for intervention description and replication (TIDieR) checklist and guide. BMJ 2014;348:g1687. [Crossref] [PubMed]
42. Lei H, Xin H, Guan R, et al. Low-intensity Pulsed Ultrasound Improves Erectile Function in Streptozotocin-induced Type I Diabetic Rats. Urology 2015;86:1241.e11-8. [Crossref] [PubMed]
43. Liu S, Jiang C, Hu J, et al. Low-Intensity Pulsed Ultrasound Enhanced Adipose-Derived Stem Cell-Mediated Angiogenesis in the Treatment of Diabetic Erectile Dysfunction through the Piezo-ERK-VEGF Axis. Stem Cells Int 2022;2022:6202842. [Crossref] [PubMed]
44. Chen Y, Xiao M, Zhao L, et al. Low-Intensity Pulsed Ultrasound Counteracts Advanced Glycation End Products-Induced Corpus Cavernosal Endothelial Cell Dysfunction via Activating Mitophagy. Int J Mol Sci 2022;23:14887. [Crossref] [PubMed]
45. Peng DY, Reed-Maldonado AB, Lin GT, et al. Low-intensity pulsed ultrasound for regenerating peripheral nerves: potential for penile nerve. Asian J Androl 2020;22:335-41. [Crossref] [PubMed]
46. Liu SY, Liu SY, Han BM, et al. Low-intensity pulsed ultrasound treatment in erectile dysfunction. Asian J Androl 2025;27:673-9. [Crossref] [PubMed]
47. Li Z, Zhou Z, Chen H, et al. Advances in the clinical applications and research progress of low-intensity pulsed ultrasound for erectile dysfunction: a narrative review. Sex Med Rev 2025;13:623-31. [Crossref] [PubMed]
48. Zhang Q, Lv B, Yu H, et al. Evaluation of tadalafil in combination with Low-Intensity Pulsed Ultrasound (LIPUS) for erectile dysfunction: a protocol for a randomized controlled trial (SPIRIT Compliant). Trials 2025;26:591. [Crossref] [PubMed]
49. Kim S, Cho MC, Cho SY, et al. Novel Emerging Therapies for Erectile Dysfunction. World J Mens Health 2021;39:48-64. [Crossref] [PubMed]