Efficacy and Tolerability Outcomes of Minimally Invasive Surgical Treatments for Benign Prostatic Hyperplasia: A Random-Effects Network Meta-Analysis

Abstract

Purpose

Minimally invasive surgical treatments (MIST) provide emerging alternatives to transurethral resection of the prostate (TURP) for benign prostatic hyperplasia (BPH). This network meta-analysis provides a comparative analysis of clinical efficacy and tolerability for individual MIST versus TURP in BPH.

Materials and Methods

A PRISMA-compliant systematic review was performed by searching the PubMed, EMBASE, and Cochrane Library databases from inception to 31 August 2023. Eligible studies compared post-treatment change scores or relative risks between ≥2 distinct BPH interventions for International Prostate Symptom Score (IPSS), peak urinary flow rate (Qmax), post-void residual urine, Sexual Health Inventory in Men (SHIM) and Men’s Sexual Health Questionnaire (MSHQ) scores, and re-intervention rate. Pairwise comparisons were synthesized through a random-effects frequentist model to determine aggregate estimates for treatment-outcome associations.

Results

Of 4,416 screened articles, 59 studies were included for quantitative analysis, comprising 21,078 patients across 47 unique cohorts. Prostatic arterial embolization (PAE) and Aquablation appeared to induce similar reductions in IPSS to TURP; however, there was insufficient data to evaluate the tolerability of PAE or Aquablation. Water vapor thermal therapy (WVTT) and prostatic urethral lift (PUL) were less efficacious at reducing IPSS than TURP but exhibited greater preservation of SHIM and MSHQ scores. TURP demonstrated the lowest relative risk of re-intervention.

Conclusions

PAE and Aquablation appear to be efficacious approaches to BPH management, with PAE also producing a comparable increase in Qmax to TURP. WVTT and PUL appear less effective than TURP but are associated with lower risk of male sexual dysfunction.

INTRODUCTION

Transurethral resection of the prostate (TURP) is accepted as the standard surgical care for lower urinary tract symptoms (LUTS) secondary to benign prostatic hyperplasia (BPH) [1]. However, TURP is invasive and associated with morbidity, including sexual dysfunction [2,3,4]. Hence, many patients are concerned about TURP-related complications, especially with the increasing global BPH prevalence [5].

In recent years, contemporary surgical interventions have seen the advent of minimally invasive surgical treatments (MIST) for BPH/male LUTS such as prostatic arterial embolization (PAE), thermal ablation, mechanical stents, and selective laser therapy [6,7]. New devices such as Aquablation incorporate robot-assisted minimally invasive technology for prostate tissue resection [7]. Relative to TURP, these MISTs have reported considerably lower postoperative complications [2,6] and in many instances, can be performed without general anesthesia in an outpatient setting, thereby increasing surgical accessibility and decreasing the overall treatment cost. Each MIST has distinct therapeutic applications, efficacy and safety profiles, and head-to-head evidence on the efficacy of individual MIST versus TURP remains controversial due to study methodology and unclear outcomes [8,9,10,11]. While many trials report non-inferiority to TURP [12,13], several studies indicate lower relative surgical efficacy for MIST [14,15], and TURP remains the definitive procedure for BPH/LUTS refractory to previous MIST and longer-term success [16].

The following systematic review and network meta-analysis (SRNMA) aims to compare longitudinal efficacy and adverse effects outcomes for patients receiving MIST versus TURP for BPH.

MATERIALS AND METHODS

1. Literature search

This prospectively registered SRNMA was performed according to the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) guidelines (PROSPERO ID: CRD42023449722). A three-pronged search strategy was designed using key terms pertinent to BPH (e.g. “benign prostatic hyperplasia”, “lower urinary tract symptoms”, surgical/medical intervention (e.g. “transurethral resection of the prostate”, “prostatic arterial embolization”, “prostate surgery”), and clinical outcomes (e.g. “haematuria”, “re-intervention”, “International Prostate Symptom Score”). Database searches were executed on PubMed, EMBASE, and Cochrane Library from inception to 31 August 2023. Manual screening of reference lists was conducted for comprehensiveness by two of the authors (BNHS and ST) with the senior authors (EC and HR) acting as final arbitrators.

2. Eligibility criteria

Eligible studies compared at least two distinct interventions for BPH, including at least one surgical intervention, in men aged ≥40 years with LUTS. The minimum cell size for inclusion was 10. We included both randomized controlled studies (RCTs) and non-randomized cohort studies. We excluded cohort studies without a comparator; studies including patients with secondary prostate or urinary tract co-morbidities; non-primary articles; case reports and case series, except for those that met the cell size criterion; and in vitro studies.

3. Screening and data extraction

Duplicate detection was performed in EndNote 20 (Clarivate). Abstracts and full-text screening were independently performed by two reviewers (BNHS, RM) through the digital platform Rayyan (Rayyan), with discrepancies arbitrated by two consultant urologists (HR, EC). Quantitative data was extracted and cross-validated by two independent reviewers (BNHS, RM).

4. Outcome measures

The primary measures for efficacy included change scores for the International Prostate Symptom Score (IPSS), flow rate (Qmax), and post-void residual urine (PVR). Adverse effect outcome measures included the re-intervention rate and the Sexual Health Inventory in Men score (SHIM). Outcomes were collected as the most contemporaneous paired outcome measure for each study to a maximum of 60 months post-intervention.

Data on additional outcomes were collected but did not demonstrate a sufficiently robust comparative network to draw strong conclusions; these included the International Index of Erectile Function 15-item (IIEF-15), the Male Sexual Health Questionnaire Short Form scales (MSHQ-EjD and MSHQ-Bother); and categorical rates of urinary tract narrowing, ejaculatory disorders, erectile dysfunction, and chronic urinary incontinence.

The risk of bias for each study was assessed using either the Cochrane Risk of Bias Tool for Randomized Trials (RoB-2) or the Risk of Bias in Non-Randomized Studies tool (ROBINS-I).

5. Missing data

In instances where outcome measures were solely reported as medians and interquartile ranges, we estimated medians as means and imputed standard deviations (SDs) as interquartile ranges divided by 1.35 [17]. Absent change SDs were computed using correlation coefficients derived from corresponding studies for that outcome [18].

6. Statistical analysis

Statistical analyses were performed in R 4.3.1 with the netmeta 2.8.2 package. Cohort characteristics were reported as mean±SD. Network meta-analysis was conducted through a frequentist random-effects model applying the Dersimonian-Laird method for weighting. Direct pairwise comparisons were connected to generate indirect comparisons for each included intervention.

Effect sizes for continuous outcomes were normalized according to the reference intervention and reported as absolute mean differences (MDs) from baseline. To determine aggregate effect sizes for continuous outcomes, MDs from baseline and 95% confidence intervals (95% CIs) for the reference intervention were estimated using a pooled random-effects meta-analysis specific to that intervention. TURP was used as a reference intervention unless otherwise specified.

Categorical outcomes were reported as relative risks (RRs); pairwise comparisons lacking events in one or both arms were excluded. To establish a treatment hierarchy for each outcome measure, we performed 200,000 frequentist simulations per outcome model to estimate mean rank; we subsequently employed these to perform Surface Under the Cumulative Ranking Curve (SUCRA) scoring, for which values of 0% indicate absolute inferior therapies and values of 100% indicate superior therapies [19].

For each outcome, comparative structures were visualized through network diagrams, for which point and arc thickness correspond to the number of intervention-specific studies and comparisons, respectively. Heterogeneity was quantified through the I² statistics and evaluated using Cochran’s Q. Funnel plots and Egger’s test were performed to identify the risk of publication bias. Heterogeneity across studies was assessed 117 using the I² statistic with I² >50% considered moderate-to-high heterogeneity.

7. Sensitivity analyses

Primary studies on BPH management exhibit considerable heterogeneity in follow-up duration and study design [9,20]. To explore how these factors influenced our overall findings, we performed two independent sensitivity analyses. Firstly, we evaluated networks restricted to short (≤12 months post-intervention) and long-term data capture (>12 months post-intervention). Studies that provided data across both short- and long-term timeframes were included in both networks, with the short-term model censored at the latest follow-up point within 12 months of the procedure. Secondly, to investigate the impact of participant selection, we constructed distinct networks to separate RCTs and non-randomized cohort studies (non-RCTs). We compared these time- and randomization-specific models to our primary results to identify potential sources of nuance or bias introduced by between-study variability.

RESULTS

1. Study and participants' characteristics

Database search retrieved 4,416 records (PubMed: 1,020; Embase: 2,103; Cochrane Library: 1,293), from which 2,664 non-duplicate abstracts and 80 full-text articles were screened. Following the application of eligibility criteria, we included 59 articles describing 47 independent cohorts (Fig. 1) [12,13,14,15,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63]. Of these, 35 were RCTs [12,13,14,15,21,25,26,28,29,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,51,52,55,56,57,58,59] and 12 were non-randomized cohorts [22,23,24,27,30,50,53,54,60,61,62,63]. 41 study cohorts were prospective [12,13,14,15,21,22,25,26,28,29,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,55,56,57,58,59,60,61,62,63], while 6 were retrospective [23,24,27,30,53,54]. Baseline characteristics for the 21,078 included participants were aggregated and stratified by intervention (Table 1).

Fig. 1. PRISMA flow diagram.

Table 1. Pooled characteristics of eligible studies stratified by intervention.

Intervention	Cohort	Participant	Age (y)	PVR (mL)	IPSS	AUA	PSA (ng/mL)
TURP	23	12,176	68.8±9.5	43.1±18.8	21.4±6.3	21.3±6.4	1.4±2.5
PUL	11	4,281	67.3±9.4	40.7±10.8	21.2±5.7	21.2±5.7	1.2±1.6
TUMT	18	1,158	64.7±17.6	41.9±19.3	20.3±7.5	18.5±7.2	1.3±1.5
PAE	11	732	68.2±8.9	40.3±40.3	22.2±5.9	21.6±5.8	3.6±3.9
Sham	14	714	64.6±6.4	45.1±15.2	23.5±6.7	20.4±6.3	1.2±1.5
WVTT	6	667	64.8±8.5	45.4±12.9	19.0±6.1	18.7±6.1	2.0±2.0
TUNA	8	620	66.2±8.7	30.4±10.5	23.1±4.0	25.1±6.4	0.6±1.2
Med	3	465	62.8±6.7	43.3±21.3	20.9±4.2	20.9±4.3	2.4±2.2
TIND	1	128	61.5±6.5	43.4±15.5	22.1±6.8	22.1±6.8	2.2±2.3
Aquablation	1	117	66.0±7.3	54.1±16.2	22.9±6.0	22.9±6.0	3.7±3.3
US-TPLA	1	20	73.9±9.2	70.7±23.8	22.7±5.2	22.7±5.5	NR
All	47	21,078	67.7±9.8	44.7±19.3	21.3±6.2	20.6±6.2	1.7±2.1

Values are presented as mean±standard deviation.

PVR: post-void residual urine, IPSS: International Prostate Symptom Score, AUA: American Urological Association; PSA: prostate specific antigen, TURP: transurethral resection of the prostate, PUL: prostatic urethral lift, TUMT: transurethral microwave thermotherapy, PAE: prostatic arterial embolization, WVTT: water vapor thermal therapy, TUNA: transurethral needle ablation, TIND: temporarily implanted nitinol device system, US-TPLA: ultrasound-guided transperineal laser ablation.

A total of 11 interventions were reviewed, including 9 surgical procedures: TURP (monopolar or bipolar, n=23) [1,2,3,6,7,8,9,11,12,13,14,15,17,18,19,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65]; prostatic urethral lift (PUL, n=11) [13,24,28,30,41,46,53,54,55,58,61]; transurethral microwave thermotherapy (TUMT, n=18) [15,21,22,25,33,34,35,36,37,38,39,45,48,49,50,53,57,62]; PAE (n=11) [14,24,27,29,40,44,51,52,54,59,63], convective radiofrequency water vapor thermal therapy (WVTT, n=6) [23,24,30,42,47,61]; transurethral needle ablation (TUNA, n=8) [22,26,32,43,50,53,56,60], temporarily implanted nitinol device system (TIND, n=1) [31], Aquablation (n=1) [12], and ultrasound-guided transperineal laser ablation (US-TPLA, n=1) [27]. To facilitate head-to-head comparisons, we also included studies that employed pharmacological therapy, including mono or combined therapy with 5-alpha reductase inhibitors or alpha-1 blockers (Med, n=3) [37,42,59]. Lastly, sham intervention (n=14) was considered as a negative control [21,25,28,31,35,36,46,47,48,49,51,55,57,58].

2. Efficacy outcomes

Thirty studies assessing 11 interventions reported change scores for IPSS (Fig. 2) [12,13,14,15,22,24,25,26,27,28,29,30,31,32,33,39,41,42,43,45,47,48,51,55,58,59,60,61,62,63]. Overall, Aquablation (MD: -14.9; 95% CI: -18.5, -11.3), TURP (MD: -13.0; 95% CI: -14.2, -11.8), and PAE (MD: -13.0; 95% CI: -15.0, -10.9) demonstrated the greatest post-intervention reductions in IPSS relative to the remaining interventions, though only one study assessed Aquablation. TIND (MD: -6.34; 95% CI: -10.6, -2.10) and sham control (MD: -3.84; 95% CI: -5.94, -1.74) produced the lowest aggregate reduction in IPSS. There was significant outcome heterogeneity between trials assessing IPSS (I²=93.1%, Cochran’s Q p<0.001), but no evidence of significant publication bias (Egger’s p=0.093).

Fig. 2. Results of frequentist random-effects network meta-analysis evaluating the effect of eligible interventions on change scores for International Prostate Symptom Score (IPSS). (A) Network diagram, indicating characteristics and connectedness of included studies; points indicate the number of studies assessing each intervention, while arcs represent direct comparisons between interventions. (B) Forest plot depicting estimated change scores for each intervention. Transurethral resection of the prostate (TURP) is applied as a reference, with an effect size computed using pooled meta-analysis specific to TURP. Treatment hierarchies are established at right using 200,000 frequentist simulations to estimate mean rank and surface under the cumulative ranking curve (SUCRA) score. 95% CI: 95% confidence interval, Mean diff: mean difference from baseline, PAE: prostatic arterial embolization, PUL: prostatic urethral lift, TIND: temporarily implanted nitinol device system, TUMT: transurethral microwave thermotherapy, TUNA: transurethral needle ablation, US-TPLA: ultrasound-guided transperineal laser ablation.

Twenty-eight studies including 11 interventions, evaluated change scores for maximum urinary flow rate (Qmax) (Fig. 3) [12,13,15,22,25,26,27,28,29,30,31,32,33,34,35,39,41,42,43,45,46,47,48,49,55,58,59,60]. At the last follow-up, Aquablation (MD: 12.8; 95% CI: 9.27, 16.3) and TURP (MD: 10.5; 95% CI: 8.82, 12.20) demonstrated the highest estimated improvements in Qmax across the eligible interventions; however, as with IPSS, just one study assessed Aquablation. TUMT (MD: 4.17; 95% CI: 2.73, 5.61) and TUNA (MD: 3.50; 95% CI: 1.95, 5.05) exhibited moderate aggregate improvement in Qmax; all other interventions showed low or no improvement. There was significant heterogeneity between trials assessing Qmax (I²=90.3%; Cochran’s Q p<0.001), but no evidence of publication bias (Egger’s p=0.680).

Fig. 3. Results of frequentist random-effects network meta-analysis evaluating the effect of eligible interventions on change scores for maximum urinary flow rate (Qmax). Both (A) Network diagram and (B) forest plot are constructed as per Fig. 2. 95% CI: 95% confidence interval, Mean diff: mean difference from baseline, SUCRA: Surface under cumulative ranking curve score, PAE: prostatic arterial embolization, PUL: prostatic urethral lift, TIND: temporarily implanted nitinol device system, TUMT: transurethral microwave thermotherapy, TUNA: transurethral needle ablation, US-TPLA: ultrasound-guided transperineal laser ablation, TURP: transurethral resection of the prostate.

Twenty-five studies assessing 11 interventions reported change scores for PVR (Fig. 4) [12,13,15,22,25,26,27,29,30,31,32,33,34,39,41,42,43,45,47,48,49,55,58,59,60]. TURP (MD: -57.4; 95% CI: -64.2, -50.6), PAE (MD: -60.6; 95% CI: -116, -5.5), and US-TPLA (MD: -53.4; 95% CI: -133, 26.5) demonstrated the highest estimated reductions in PVR post-intervention, though only one study assessed US-TPLA. The remaining interventions exhibited a less favorable effect on PVR; in particular, both TIND (MD: 12.0; 95% CI: -51.2, 75.3) and sham control (MD: 19.2; 95% CI: -13.3, 51.7) resulted in an estimated increase in PVR following an intervention. However, the between-trial heterogeneity was high (I²=96.2%; Cochran’s Q p<0.001), resulting in low confidence for estimated effect sizes. Publication bias was not evident (Egger’s p=0.742).

Fig. 4. Results of frequentist random-effects network meta-analysis evaluating the effect of eligible interventions on change scores for post-void residual (PVR). Both (A) Network diagram and (B) forest plot are constructed as per Fig. 2. 95% CI: 95% confidence interval, Mean diff: mean difference from baseline, SUCRA: surface under cumulative ranking curve score, PAE: prostatic arterial embolization, PUL: prostatic urethral lift, TIND: temporarily implanted nitinol device system, TUMT: transurethral microwave thermotherapy, TUNA: transurethral needle ablation, US-TPLA: ultrasound-guided transperineal laser ablation, TURP: transurethral resection of the prostate.

3. Adverse effect outcomes

Ten studies across 7 interventions investigated change scores in SHIM (Fig. 5) [12,13,24,28,29,31,41,46,55,58]. TIND exhibited a moderate increase in post-intervention SHIM (MD: 2.64; 95% CI: 0.25, 5.02). Sham control, WVTT, and PUL demonstrated mild increases in SHIM, while Aquablation and TURP were associated with negligible change scores. Contrastingly, PAE was associated with a high estimated reduction in SHIM (MD: -4.96; 95% CI: -8.20, -1.71), although this intervention was assessed in only one study. There was low heterogeneity in studies assessing SHIM (I²=27.0%; Cochran’s Q p=0.242) and no evidence of publication bias (Egger’s p=0.279).

Fig. 5. Results of frequentist random-effects network meta-analysis for eligible studies on change scores for Sexual Health Inventory in Men score (SHIM). Both (A) Network diagram and (B) forest plot are constructed as per Fig. 2. 95% CI: 95% confidence interval, Mean diff: mean difference from baseline, SUCRA: surface under cumulative ranking curve score, PAE: prostatic arterial embolization, PUL: prostatic urethral lift, TIND: temporarily implanted nitinol device system, TUMT: transurethral microwave thermotherapy, TUNA: transurethral needle ablation, US-TPLA: ultrasound-guided transperineal laser ablation, TURP: transurethral resection of the prostate.

Nine interventions across 14 studies assessed the incidence of re-intervention at the last follow-up (Fig. 6) [12,13,23,24,30,35,36,37,38,40,41,43,59,61]. Although all therapies demonstrated a higher RR for re-intervention than TURP, the absolute incidence of re-intervention was low across studies, resulting in wide CIs for this outcome measure. There was moderate heterogeneity between trials (I²=50.1%; Cochran’s Q p=0.061) and no evidence of publication bias (Egger’s p=0.185).

Fig. 6. Results of frequentist random-effects network meta-analysis evaluating the risk of any re-intervention at last follow-up for eligible interventions. (A) Network diagram constructed as per Fig. 2. (B) Forest plot depicting estimated relative risks following each intervention; TURP is employed as a reference. Treatment hierarchies are established at right using 200,000 frequentist simulations to estimate mean rank and SUCRA score. 95% CI: 95% confidence interval, SUCRA: surface under cumulative ranking curve score, PAE: prostatic arterial embolization, PUL: prostatic urethral lift, TIND: temporarily implanted nitinol device system, TUMT: transurethral microwave thermotherapy, TUNA: transurethral needle ablation, US-TPLA: ultrasound-guided transperineal laser ablation, TURP: transurethral resection of the prostate.

Data on additional outcome measures was sparsely reported. There was insufficient evidence to estimate relative efficacy concerning change scores for IIEF-15. TURP data indicated a moderate decrease in post-intervention MSHQ-EjD scores (MD: -3.04; 95% CI: -3.56, -2.52), as compared to minimal change for PUL, WVTT, Aquablation, and sham control. This was corroborated by the MSHQ-Bother model, in which WVTT (MD: -0.65; 95% CI: -1.63, 0.32), PUL (MD: -0.54; 95% CI: -1.09, 0.02) and sham control were each associated with reduced change scores relative to TURP (MD: -0.04; 95% CI: -0.33, 0.26). For binary event outcomes, MIST demonstrated a strong trend towards lower RR versus TURP for post-procedure urinary tract narrowing, ejaculatory disorders, erectile dysfunction, and urinary incontinence. However, the absolute incidence of each of these complications was low.

4. Sensitivity analyses

To evaluate whether outcomes were rapidly achieved and/or durable, we performed a sensitivity analysis in which treatment models were restricted to either short (≤12 months post-intervention) or long-term (>12 months post-intervention) measures and compared to the primary treatment models (Fig. 2, 3, 4, 5, 6). No long-term treatment models had available data on TIND and US-TPLA, and the Qmax, PVR, SHIM, and re-intervention treatment models also lacked comparative data on WVTT, PAE, and pharmacological therapy. As in the aggregate treatment models, TURP and Aquablation continued to demonstrate the highest relative improvements in change scores for IPSS and Qmax across both follow-up durations. Although PUL was associated with moderate improvements in IPSS (MD: -11.3; 95% CI: -13.6, -8.97) and Qmax (MD: 2.65; 95% CI: 0.80, 4.50) in the short-term model, both of these effects were attenuated in the long-term model (IPSS MD: -5.77; 95% CI: -9.30, -2.25; Qmax MD: -2.41; 95% CI: -8.36, 3.54). We note that only one study assessed PUL at >12 months post-intervention.

Consistent with the aggregate PVR model, TURP (MD: -54.6; 95% CI: -65.7, -43.5), PAE (MD: -66.5; 95% CI: -114, -19.5), and US-TPLA (MD: -66.5; 95% CI: -114, -19.5) continued to demonstrate high post-procedure reductions in the short-term PVR model. However, no studies evaluated PAE or US-TPLA in the long-term PVR model. TUMT was associated with a greater post-procedure reduction in long-term PVR (MD: -75.4; 95% CI: -125, -26.4) than TURP (MD: -50.2; 95% CI: -67.7, -32.8), although heterogeneity was high (I²=81.6%, Cochran’s Q p<0.001) and only two studies assessed PVR at ≥12 months following TUMT. For SHIM, TIND demonstrated a higher average change score than TURP in both the aggregate and short-term models (MD: 2.71; 95% CI: 0.41, 5.00), but was not assessed in the long-term model. Corroborating the aggregate re-intervention model, TURP remained the procedure with the lowest risk of re-intervention in both the short- and long-term models.

We conducted a further sensitivity analysis delineating between RCTs and non-randomized cohort studies (non-RCTs). For the aggregate model, the RCT-specific model indicated concordant effect sizes and treatment hierarchies for IPSS, Qmax, PVR, SHIM, and re-intervention. There were insufficient non-RCTs to estimate robust effect sizes for this subgroup.

5. Risk of bias

The majority of the 35 included RCTs appropriately mitigated against the risk of bias, and 4 (11%) demonstrated a serious risk of bias in any category. Of the 12 included non-RCTs, 4 (33%) were at serious risk of bias, particularly due to imbalances in patient selection or insufficient consideration of confounding domains.

DISCUSSION

This network meta-analysis synthesized evidence from 47 cohorts across 59 publications, comprising 21,078 participants who underwent treatment for BPH. To develop a more concrete estimate of post-intervention effects for continuous outcomes, we first generated a pooled random-effects meta-analysis specific to TURP and used this estimate as a reference category for all other interventions.

TURP was associated with improved aggregate scores in terms of IPSS, Qmax, and PVR outcomes. Both PAE and Aquablation demonstrated comparable improvement in IPSS to TURP, although Aquablation was only assessed in one study. TUNA, WVTT, TUMT, US-TPLA, PUL, and TIND were associated with reduced improvement in change scores for IPSS relative to TURP; however, except for TIND, the CIs for each of these MIST overlapped with that of TURP. Aquablation and TURP produced the highest increases in post-intervention Qmax, while PAE, TURP, and US-TPLA resulted in the highest reduction in post-intervention PVR.

TURP, WVTT, PUL, and Aquablation resulted in a minimal aggregate change in SHIM post-intervention in tolerability outcomes. Evidence from one study suggested a greater improvement in SHIM score for TIND relative to TURP, and evidence from another study suggested a lower improvement in SHIM for PAE relative to TURP; however, this data is insufficient to conclude the effect of these specific interventions on erectile dysfunction. TURP was associated with a high reduction in post-intervention MSHQ-EjD scores; in contrast, WVTT, PUL, and Aquablation appeared to induce minimal effects, with comparable change scores to sham control. Both WVTT and PUL also demonstrated a reduction in MSHQ-Bother scores relative to TURP. Nonetheless, TURP demonstrated the lowest RR of re-intervention, followed by TUMT and Aquablation; pharmacological therapy, WVTT, and TUNA were associated with the greatest RR of re-intervention.

We performed sensitivity analyses of treatment models restricted to both short (≤12 months) and long-term follow-up (>12 months) following the intervention to assess surgical durability. The above findings largely persisted in both the short- and long-term models, with TURP and Aquablation remaining highly ranked for change scores in IPSS across all follow-up durations. No included studies assessed outcomes in TIND and US-TPLA at >12 months post-intervention, and results from a single study suggested lower durability for PUL on IPSS and Qmax in the long-term model. The relatively sparse data on several modalities such as Aquablation and US-TPLA limits the generalizability and actual analysis of surgical durability in comparison with established MIST such as PUL. A recent SRNMA found that the retreatment rate with these new surgical interventions is comparable if not better than TURP, in the short-term, although the cumulative reintervention rate also varied depending on the types of new surgical interventions performed [20].

Notably, MIST in this series appeared to have disparate efficacy profiles. For example, PAE was found to be similarly effective to TURP at reducing IPSS and PVR but produced only a minimal increase in Qmax. Similarly, WVTT was estimated to induce a considerable reduction in IPSS while only producing mild improvements in Qmax and PVR. Although this is potentially due to the high intra-cohort heterogeneity observed in this series, an alternative inference is that patient-reported symptom scores may not necessarily converge with conventional urodynamic measures. Previous cohort studies have identified only moderate correlations between IPSS and Qmax/PVR [64,65,66,67], suggesting the importance of targeting therapy to the specific management goals of each patient. In this series, we note that while Aquablation and PAE produced similar improvements in IPSS to TURP, Aquablation was associated with high improvement in Qmax but not PVR, while PAE was associated with high improvement in PVR but not Qmax. Therefore, PAE may potentially be a preferable MIST for patients with chronic retention, although prospective controlled trials would be required to further evaluate this hypothesis.

Non-randomized cohort studies represented 12 of the 47 cohorts in this meta-analysis. While non-randomized studies are more susceptible to bias than RCTs, they also involve a broader range of participants and are generally able to sustain a longer duration of follow-up [68]. This is particularly pertinent for this network meta-analysis given that RCTs tend to recruit participants who are younger, less comorbid, and at lower risk of adverse events than patients in the general population [69,70]. As the reported advantage of MIST over TURP is a lower incidence of postoperative complications [2,6], restricting our analysis to RCTs may preclude treatment comparisons between patients in high-risk subgroups. Therefore, we elected to include non-randomized studies to expand sample size and patient representation. A secondary benefit of this was improved network connectedness, and we were thus able to integrate head-to-head data on US-TPLA versus PAE [27] and WVTT versus PUL [24,30], among others. Although we could not estimate confident effect sizes for the non-RCT model, the concordance between the aggregate and RCT models in our sensitivity analysis is encouraging and suggests that the non-RCT studies provided additional data without skewing the direction or effect sizes of the primary model.

By integrating a broader range of studies, interventional modalities, and follow-up durations, our network meta-analysis expands upon several previous meta-analyses in the field of BPH management [8,9,71,72]. Tanneru et al [71] investigated comparative data on Aquablation, WVTT, PUL, and TURP for patients with a prostate size <80 mL, but included only 4 studies. Franco et al [8] performed a comprehensive meta-analysis of 27 RCT cohorts with head-to-head data on 3,017 participants. As with our findings, Franco et al observed that TURP demonstrated the highest likelihood of efficacy for IPSS and avoiding re-intervention; however, Qmax and PVR were not assessed, and TUNA, Aquablation, pharmacological therapy, and US-TPLA were omitted as comparators.

A recent meta-analysis by Cornu et al [9] investigated comparative BPH management in 63 RCT cohorts at 6-month and 12-month post-intervention timepoints. Notably, 34 of these studies exclusively compared monopolar versus bipolar TURP, while 3 studies compared PAE using different bead sizes. The remaining 26 studies spanned six surgical modalities and sham control but omitted TUNA, Aquablation, pharmacological therapy, and US-TPLA. In contrast to the present study, durability was not assessed beyond 12 months post-intervention. We note also that the meta-analysis by Cornu et al [9] was funded and authored by Boston Scientific, the manufacturer of the WVTT system.

These findings should be interpreted with caution. Firstly, there was limited head-to-head data on several modalities, with TIND [31], Aquablation [12], and US-TPLA [27] only represented by a single comparative trial each. This was compounded by heterogeneous outcomes reporting and adverse event measures beyond studies. For example, erectile dysfunction was reported through several distinct methods: a continuous SHIM score [13], a patient-reported binary outcome [55], or a binary outcome defined by a post-intervention decrease in SHIM score of ≥6 points [12]. As these approaches were not numerically comparable, treatment networks for complications such as urinary tract stricture and incontinence were poorly connected and inconclusive.

Current published guidelines in male LUTS advocate for a stepwise strategy from conservative measures with medical therapy to surgery as last resort [73,74]. However, advances in BPH technology coupled by greater demands from patients wanting a more definitive treatment with excellent clinical efficacy and safety profile, has led to the adoption of various innovative minimally invasive BPH technology that improves voiding parameters and preserving sexual function. Nonetheless, available literature on cost-effective analysis is limited [75] and this could be related to various factors including the actual and perceived values clinician or patient place on the a (single) surgical intervention instead of daily medications. Clinicians should provide a holistic approach and shared decision making with patients on the benefits and risks of various BPH therapeutic approaches based on best-evidenced clinical practice. The emerging role of metabolic milieu, genetic predispositions, diagnostic evaluation with novel predictive modellings, and prevention medicine require clarification and fills the current gaps in BPH literature. Further validation with objective assessment in large well-designed and multi-center studies is desired, to examine the actual role of MIST in clinical algorithm and importantly, if MIST can be considered the new standard of care.

1. SWOT analysis

A SWOT analysis (which stands for strengths [S], weaknesses [W], opportunities [O], and threats [T]) condenses the critical points of our SRNMA. While the external elements "opportunities" (O), "threats" (T), and "strengths" (S) relate to the internal environment, respectively, the external factors "weaknesses" (W) and "strengths" (S) relate to the internal environment [76]. Using the SWOT method, we highlight the key benefits and limitations of our SRNMA based on SWOT, comparing MIST and TURP in BPH surgery.

1) Strengths:

• 1. Inclusion of the largest studies with the largest case numbers in an SRNMA assessing MIST and TURP in BPH surgery

  2. The studies included in this meta-analysis had superior patient selection and study design quality

  3. Comprehensive data analysis on MIST and TURP in BPH surgery involving 47 study cohorts

  4. There were highly significant differences in the outcomes of patients between MIST and TURP in BPH surgery

2) Weaknesses:

• 1. A limited number of direct comparative studies among various MIST

  2. The relatively short follow-up time with most papers reporting outcomes of less than 24 months

  3. Sparse data on new therapies such as Aquablation, US-TPLA and OptiLume system

  4. Included studies still exhibited significant heterogeneity with lack of uniformity in reporting sexual function and adverse outcomes reduces interpretability and mixed outcomes

  5. Publication bias is possible in this meta-analysis

3) Opportunities:

• 1. To help clinicians better comprehend the outcomes between MIST and TURP in BPH surgery

  2. To serve researchers in conducting more standardized research methodology on the outcomes between MIST and TURP in BPH surgery

4) Threats:

• 1. Lack of direct comparative studies among the various MISTs and against TURP may lead to a lack of sufficient RCTs to validate the findings

  2. Failure to consider this study may lead to an underestimation of the actual surgical outcomes in terms of clinical efficacy, durability and safety between MIST and TURP in BPH surgery

CONCLUSIONS

Moderate-quality evidence from 47 study cohorts indicates that TURP remains the definitive procedure for BPH, with highly ranked improvements in IPSS, Qmax, and PVR and the lowest RR of re-intervention among all included interventions. Both WVTT and PUL were associated with higher post-intervention SHIM and MSHQ-EjD scores than TURP and appeared to be more tolerable. Further head-to-head studies and uniformity in reporting standards are required to provide better evidence-based surgical guidance in the current landscape of MIST in BPH/LUTS.