Synthetic sling or artificial urinary sphincter for men with urodynamic stress incontinence after prostate surgery: the MASTER non-inferiority RCT

Men undergoing prostate surgery, particularly radical prostatectomy for cancer or TURP for benign disease, frequently experience stress urinary incontinence (SUI) that can persist despite conservative measures. AUS implantation is recommended for troublesome SUI persisting >12 months post-surgery but is invasive, costly, requires manual dexterity, and may need revision. Male synthetic slings, typically non-adjustable transobturator polypropylene mesh, have been increasingly used as a potentially less invasive, easier-to-operate and less costly alternative. However, prior to this study there were no adequately powered RCTs comparing sling with AUS, and guidance (NICE CG97) called for RCT evidence. MASTER aimed to determine whether the male sling is non-inferior to AUS for continence at 12 months, and to evaluate comparative harms, costs, cost-effectiveness and patient satisfaction.

Prior evidence consisted largely of case series with variable definitions of cure/improvement and limited use of patient-reported outcomes. A 2014 Cochrane review found no adequately powered RCTs comparing male sling and AUS, identifying only one small, poor-quality RCT comparing AUS versus injectable bulking agents, which favoured AUS. Guidelines (NICE CG97; EAU) recommended AUS as standard for persistent male SUI and noted that male slings should be used within trials. Observational reports suggested adjustable slings might incur more additional surgery/complications with no clear efficacy advantage over passive slings. The prevalence and persistence of post-prostatectomy incontinence are substantial, and recovery plateaus by 12 months, underscoring the need for surgical comparative effectiveness data.

Design: Multicentre, pragmatic, non-inferiority RCT with parallel NRC and embedded qualitative programme. Centres: 28 NHS urology centres in the UK. Participants: Men with urodynamic stress incontinence (USI) after radical prostatectomy or TURP, suitable for surgery. Exclusions: prior sling/AUS, unresolved bladder neck contracture or urethral stricture, insufficient dexterity for AUS, inability to consent/complete documentation; after NRC closure, men unwilling to be randomised were excluded from RCT. Randomisation: 1:1 to male sling or AUS via CHaRT web/IVR system; minimised on previous prostate surgery type (radical prostatectomy/TURP), prior radiotherapy (yes/no), and centre. Blinding: Not possible for participants/surgeons; 12‑month clinic assessors were asked to remain blinded where feasible. Interventions: Non-adjustable, suburethral, transobturator male sling; or standard AUS (cuff around urethra, pressure-regulating balloon, scrotal pump), deactivated 4–6 weeks post-implant then activated. Follow-up: Participant questionnaires at baseline (pre-surgery), 6 months post-surgery, and 12 and 24 months post-randomisation; 3‑day bladder diaries; 12‑month clinic with 24‑hour pad test. Up to two reminders for questionnaires. NRC followed by questionnaires/electronic data only. Outcomes: Primary clinical outcome—participant-reported continence at 12 months defined by ICIQ-UI SF Q3='never' and Q4='none' (DMC also recommended reporting a less strict definition including 'once a week or less often' and 'a small amount'). Primary economic outcome—incremental cost per QALY at 24 months (NHS/PSS perspective). Secondary outcomes—incontinence symptom scores (ICIQ-UI SF), ICIQ-MLUTS (voiding, incontinence), ICIQ-MLUTSsex, EQ-5D-3L, SF-12, pad use, 24‑hr pad weight (12 months), PGI-I, satisfaction, recommend surgery, return to normal activities, operative metrics, hospital stay, complications, re-operations, time to further surgery. Sample size: Simulations indicated 310 participants provide 90% power to demonstrate non-inferiority (margin 15%) assuming no true difference; allowing 15% attrition increased target to 360; 380 were randomised. Statistical analysis: ITT primary analysis using generalised linear model with clustering by centre; adjusted for prior radiotherapy and baseline 24‑hr pad weight; one-sided 2.5% non-inferiority test. Sensitivity analyses included multiple imputation and pattern-mixture models for missing data; per-protocol; subgroup by baseline pad weight (≤250 g vs >250 g). Repeated measures mixed models for continuous outcomes; ordered logistic regression for ordinal outcomes. Safety: AEs/SAEs recorded via site reports and questionnaires; unrelated AEs not recorded. Health economics: Within-trial cost-utility analysis to 24 months (NHS/PSS perspective), collecting intervention resource use (staff time, device cost, theatre, anaesthesia, LOS), follow-up resource use (primary/secondary care, medications, devices, further surgery), and participant costs (pads, private care, lost productivity) in sensitivity analyses. Unit costs from national sources; utilities from EQ‑5D‑3L (UK tariffs) and SF‑6D in sensitivity. GLMs for costs (gamma log-link); OLS for QALYs; multiple imputation for missing data; bootstrapping for uncertainty and CEACs. Long-term Markov cohort model (up to 20 years) explored extrapolated cost-effectiveness with states: well post-surgery, complications (3‑month state), failure (leading to further surgery), containment, death; transition probabilities from trial data; costs/utilities by state.

Participants: 380 RCT (190 sling; 190 AUS). Surgery performed in 93.4% (sling 180/190; AUS 175/190); adherence high (allocated intervention received in 178/180 sling; 166/175 AUS). NRC: 99 included; 85 had surgery (42 sling, 43 AUS). Primary clinical outcome (12 months): Continent (strict definition) sling 20/154 (13.0%) vs AUS 25/158 (15.8%); ITT absolute risk difference -0.034 (95% CI -0.117 to 0.048); non-inferiority p=0.003—sling non-inferior to AUS within -15% margin. Less strict definition: sling 52/154 (33.8%) vs AUS 55/158 (34.8%). At 24 months (strict): sling 15.0% vs AUS 14.9% (difference -0.6%, 95% CI -9.2 to 8.0). Secondary outcomes: Both groups improved from baseline. ICIQ-UI SF score decreased from 16.1 to 8.7 (sling) and 16.4 to 7.5 (AUS); adjusted mean difference at 12 months 1.30 (95% CI 0.11 to 2.49; p=0.032), favouring AUS (higher scores worse). ICIQ-MLUTS incontinence and voiding scores were consistently worse in the sling group across time points (e.g., incontinence 12 months difference 1.66, 95% CI 0.71 to 2.61; p=0.001; voiding 12 months difference 1.33, 95% CI 0.66 to 2.00; p<0.001). EQ‑5D‑3L indices and SF‑12 physical/mental scores showed no significant differences between groups; both improved slightly over time. PGI-I 'very much better' at 12 months: sling 39.5% vs AUS 52.1%. Satisfaction 'completely satisfied': sling 30.5% vs AUS 41.6%; sling less likely to be satisfied (OR 0.44, 95% CI 0.28 to 0.69; p<0.001) and to recommend surgery (79% vs 95%; OR 0.18, 95% CI 0.07 to 0.48; p=0.001). Pad use declined in both groups; daily pad counts slightly higher in sling. Reoperations/complications: Reoperation within 24 months: sling 20/180 (11.1%) vs AUS 4/175 (2.3%); 18 sling-to-AUS conversions; 2 sling replacements; AUS group had 3 AUS replacements and 1 explant without replacement. SAEs: sling 8 (recatheterisation prolonging stay=3; mesh erosion=3; infection/urosepsis=1; GI bleed=1) vs AUS 15 (recatheterisation=3; infection=3; erosion=2; haematoma=1; bruising/inflammation=1; retention/voiding=1; pain=1; transient hypotension=1; thrombosis=1; asthma exacerbation=1); one patient had 3 SAEs. Postoperative catheterisation >24 h more common after sling (11.1% vs 3.4%). Deaths: 6 in RCT (3 per group) and 1 in NRC, none related to study surgery. Subgroups: No evidence of interaction by baseline 24‑hr pad weight (≤250 g vs >250 g) on primary outcome. Health economics (within-trial, 24 months): Mean total NHS costs lower with sling by £2497 (95% CI -£3167 to -£1875). QALYs slightly lower with sling by 0.006 (95% CI -0.060 to 0.054), not significant. ICER suggests cost savings of £425,870 per QALY lost; CEAC indicates 99% probability of sling being cost-effective at £30,000/QALY threshold. Sensitivity analyses (HRG costs, societal costs, device price variation, discounting) preserved direction; probabilities 90–100% at £30,000/QALY. Long-term model: Over 20 years, sling costs remained lower (-£1511, 95% CI -£4597 to £5577) but QALYs were fewer (-0.133, 95% CI -0.782 to 0.488); ICER £11,385 per QALY lost; probability of cost-effectiveness 42% at £30,000/QALY. Results sensitive to assumed failure/complication probabilities and use of treatment-specific vs pooled state utilities/costs. Qualitative findings: Men and surgeons reported multifactorial preferences (device operation, invasiveness, lifestyle, expectations). Both procedures had similar perioperative experiences; successful outcomes were life-changing even without complete dryness; disappointment was substantial when unsuccessful. Clear, realistic preoperative information is essential.

The trial addresses a key evidence gap by directly comparing male sling with AUS in a robust, multicentre RCT. The primary outcome demonstrated non-inferiority of the sling to AUS at 12 months for strict continence definition, indicating that in appropriately selected men either operation improves continence. Nonetheless, multiple secondary and post hoc analyses favoured AUS—lower symptom scores, greater perceived improvement, higher satisfaction, and fewer reoperations—suggesting AUS may provide superior symptom control and patient-perceived benefit for many men, particularly with heavier baseline leakage. Safety profiles were acceptable for both, though sling recipients more often required postoperative catheterisation and reoperation (typically conversion to AUS). Quality-of-life improvements were similar in generic measures. From an economic perspective over 24 months, slings were substantially less costly, primarily due to lower device and theatre costs, with only a very small and statistically non-significant decrement in QALYs, yielding a high probability of cost-effectiveness at conventional thresholds. However, long-term modelling suggests the cost-effectiveness advantage diminishes as QALY differences accumulate over time and reoperation/complication risks unfold, with considerable uncertainty due to limited long-term data—particularly for slings. Overall, the findings support shared decision-making: both options improve continence and quality of life; AUS may achieve better symptom outcomes and satisfaction but entails higher costs and device operation; slings avoid pump manipulation, are less costly, and may suit men prioritising these features, accepting a higher chance of subsequent AUS. Clear counselling on realistic expectations (improvement rather than complete dryness) and the likelihood of further surgery, especially after sling, is essential.

MASTER provides high-quality comparative evidence for surgical management of male SUI after prostate surgery. The male sling was non-inferior to AUS for participant-reported continence at 12 months, with both procedures significantly improving symptoms and quality of life and yielding high satisfaction. Secondary outcomes and reoperation rates favoured AUS overall. Within 24 months, slings were less costly with minimal QALY decrement, resulting in a high probability of cost-effectiveness; long-term cost-effectiveness is uncertain and likely attenuated. These results enable informed clinician–patient discussions tailored to patient preferences and baseline severity. Future research should include longer-term (≥5 years) follow-up of this cohort to refine estimates of durability, reoperation and complications, and incorporate more detailed pain assessment.

Follow-up limited to 24 months for primary analyses; longer-term device failures and complications may be underrepresented. No blinding of participants/surgeons; outcome assessment primarily self-reported. Pain outcomes were not captured with specific, surgical/device-related instruments. Some missing data required multiple imputation; NRC closed early due to impact on RCT recruitment. Generalisability to adjustable slings is limited as only non-adjustable transobturator slings were used. Small numbers of TURP patients limited subgroup analyses by index surgery type.