Efficacy of Vibegron in Overactive Bladder: A Bayesian Cross‐Trial Synthesis of Benefit Probabilities and Decision Thresholds

ABSTRACT

Objectives

To synthesize published placebo‐controlled trials of vibegron monotherapy for overactive bladder (OAB) and quantify posterior probabilities of benefit and decision‐threshold exceedance.

Methods

We extracted published placebo‐adjusted least‐squares mean differences at Week 12 from three randomized, double‐blind, placebo‐controlled Phase III trials evaluating vibegron monotherapy (Japan Phase III trial, Korea bridging study [50 mg], and EMPOWUR trial [75 mg]) for four voiding‐diary endpoints: micturitions/day, urgency episodes/day, urgency urinary incontinence (UUI) episodes/day, and voided volume (mL). For each endpoint, a Bayesian normal–normal random‐effects model estimated the pooled mean effect and between‐trial heterogeneity and calculated posterior probabilities of benefit and of exceeding thresholds predefined in this analysis.

Results

Pooled mean effects (95% credible intervals) favored vibegron across all four voiding‐diary endpoints: micturitions/day (−0.76, −1.33 to −0.28), urgency episodes/day (−0.69, −1.35 to −0.22), UUI episodes/day (−0.44, −0.93 to −0.03), and voided volume (+24.09 mL, 15.78–32.51). The posterior probabilities of benefit were 99.5%, 99.3%, 98.0%, and > 99.9%, respectively. The probabilities of exceeding thresholds were as follows: for micturitions/day, ≤ −0.5: 90.3% and ≤ −1.0: 14.2%; for urgency episodes/day, ≤ −0.5: 82.8% and ≤ −1.0: 12.7%; for UUI episodes/day, ≤ −0.3: 81.8% and ≤ −0.5: 35.4%; and for voided volume, ≥ +10 mL: 99.5%, ≥ +20 mL: 88.4%, and ≥ +30 mL: 5.8%.

Conclusions

Bayesian cross‐trial synthesis demonstrates a high probability that vibegron improves key OAB voiding‐diary outcomes at Week 12. By explicitly quantifying the probability of exceeding clinically interpretable decision thresholds, this approach provides a transparent assessment of both the direction and magnitude of treatment effects, facilitating clinically interpretable decision‐making.

1. Introduction

Overactive bladder (OAB) is a clinical syndrome in which urinary urgency is the core symptom and is usually accompanied by increased daytime frequency and nocturia in the absence of urinary tract infection or other obvious pathology. Urgency urinary incontinence (UUI) may be present or absent [1]. Treatment response is commonly evaluated using a patient‐completed bladder diary, which quantifies symptoms on a daily basis, such as the number of micturitions, urgency episodes, UUI episodes, and the voided volume per micturition. These diary‐based endpoints are widely employed in randomized controlled trials evaluating pharmacological therapies for OAB. In placebo‐controlled trials of OAB medications, improvement from baseline is frequently observed in both placebo and active‐treatment groups, although the magnitude of improvement varies with trial design and baseline severity [2, 3]. However, conventional reporting focused on mean differences and p values does not directly answer decision‐relevant questions such as “What is the probability that the true treatment effect is beneficial?” or “What is the probability that the benefit exceeds a clinically interpretable magnitude?” Because OAB management relies heavily on patient‐reported outcomes and treatment persistence, physician–patient communication can influence perceived effectiveness and satisfaction, underscoring the value of clinically interpretable evidence summaries [4].

Vibegron is an oral β‐3 adrenoceptor agonist that has been evaluated in multiple randomized, double‐blind, placebo‐controlled trials. In a Japanese Phase III trial [5] and a Korean bridging trial [6], vibegron 50 mg significantly improved bladder diary outcomes versus placebo at Week 12. In the international Phase III EMPOWUR trial, vibegron 75 mg similarly demonstrated placebo‐adjusted improvements in micturitions, urgency, UUI, and voided volume [7]. While these trials provide robust within‐trial efficacy estimates, they were not designed to yield a single pooled estimate that explicitly accounts for between‐trial heterogeneity, including differences in dose and region.

Bayesian evidence synthesis provides a framework to integrate results across trials while directly quantifying uncertainty in terms aligned with clinical questions. From the posterior distribution, it is possible to express results as the probability of benefit and the probability that the pooled effect exceeds a prespecified decision threshold. Importantly, these probabilities describe uncertainty in the pooled average treatment effect across trials; they do not represent the probability that an individual patient will improve.

The objective of this study was to complement conventional reports with probability‐based information to better inform clinical decision‐making. Using published aggregate results from 12‐week placebo‐controlled trials of vibegron monotherapy, we conducted a Bayesian cross‐trial synthesis that accounted for between‐trial heterogeneity to estimate pooled mean effects. These results were presented as posterior probabilities of benefit and as probabilities of exceeding clinically interpretable decision thresholds.

2. Methods

2.1. Evidence Extraction and Endpoints

We conducted a cross‐trial synthesis based on published aggregate data from three 12‐week randomized, double‐blind, placebo‐controlled Phase III trials evaluating vibegron monotherapy trials at 50 or 75 mg: the Japanese Phase III trial, the Korean bridging trial, and the international Phase III trial [5, 6, 7]. Trials with dose‐ranging designs or combination‐therapy arms were not included to maintain comparability across study populations and endpoints. For consistency across trials and endpoints, we extracted the published placebo‐adjusted least‐squares (LS) mean differences in change from baseline at Week 12 for the vibegron‐dose and placebo arms included in the present synthesis (total n = 1903; Table S1).

The following bladder diary endpoints at Week 12 were analyzed: micturitions/day, urgency episodes/day, UUI episodes/day, and voided volume per micturition (mL). In the original trial publications, these endpoints were reported as placebo‐adjusted LS mean differences with 95% confidence intervals and/or p values [5, 6, 7]. The published effect estimates were subsequently reexpressed within a Bayesian hierarchical modeling framework to derive pooled posterior distributions and probability‐based summary measures.

Ethics committee review and informed consent were not required because the study utilized previously published aggregate data and did not involve individual‐level patient data or direct participation of human subjects.

2.2. Effect Measures and Standard Errors

For continuous outcomes, the effect size y _i for trial i was defined as the published placebo‐adjusted LS mean difference (vibegron minus placebo) in change from baseline at Week 12. Standard errors (SE _i ) were derived from the reported 95% confidence intervals using SE _i  = (upper − lower)/(2 × 1.96). For episode endpoints, negative values indicate improvement (fewer episodes). For voided volume, positive values indicate improvement (larger volume).

2.3. Bayesian Random‐Effects Evidence Synthesis for Monotherapy

For each endpoint, a Bayesian normal–normal random‐effects model was fitted to synthesize results across the three monotherapy trials. Weakly informative priors were used: μ ~ Normal(0, 10²) for diary frequency endpoints and μ ~ Normal(0, 200²) for voided volume, with half‐normal priors for heterogeneity (τ ~ Half‐Normal(scale = 0.5) for diary frequency endpoints and τ ~ Half‐Normal(scale = 10) for voided volume). Posterior inference was computed via numerical integration over a τ grid of 20 001 points ranging from 0 to 10 times the prior scale, followed by Monte Carlo sampling with 200 000 posterior draws per endpoint. The pooled effect (μ), heterogeneity (τ), and all probability‐based metrics reported for the pooled effect are derived quantities from this reanalysis and were not presented in the original trial publications [5, 6, 7]. Details on the τ grid specification and Monte Carlo sampling parameters implemented in R are provided in Table S2.

2.4. Exchangeability and Interpretation

The three monotherapy trials differed in vibegron dose (50 vs. 75 mg) and in geographic setting (Japan, Korea, and multinational). Accordingly, the random‐effects model treats trial‐specific effects as exchangeable but not identical: μ should be interpreted as an average placebo‐adjusted mean effect across these heterogeneous settings, and τ quantifies between‐trial variation.

To assess robustness and influence of vibegron dose, we conducted a restricted analysis limited to trials using the data of 50 mg only. Furthermore, because inference on τ can be prior‐sensitive when only k = 3 trials are available, we assessed robustness via sensitivity analyses with alternative half‐normal τ prior scales and, to address “what should I expect in my setting?,” we additionally considered posterior predictive effects for a new trial setting (θ _new) that incorporate τ. Under the normal–normal random‐effects model y _i  ~ Normal(θ _i , SE _i ² ) and θ _i  ~ Normal(μ, τ ² ), we defined θ _new | μ, τ ~ Normal(μ, τ ² ) and summarized Pr(θ _new | data) by the posterior predictive median and 95% posterior predictive interval, using the same τ‐grid numerical integration (20 001 points) and Monte Carlo posterior sampling (200 000 draws per endpoint) as for μ and τ.

2.5. Decision‐Relevant Probability Metrics

For diary frequency endpoints, we reported the posterior probability of benefit as Pr(μ < 0). For voided volume, we reported Pr(μ > 0). We also reported the posterior probability that μ exceeded thresholds predefined in this analysis intended as clinically interpretable anchors rather than definitive minimal clinically important differences. These thresholds were informed by previously proposed minimal important differences (MIDs) for OAB diary outcomes and by clinically important change analyses from EMPOWUR [8]. Because baseline symptom burdens differ across trials and the EMPOWUR meaningful‐change definitions were percentage‐based at the patient level, these absolute thresholds should be interpreted as anchors for the pooled placebo‐adjusted mean effect rather than as responder criteria. For micturitions/day and urgency episodes/day, we evaluated thresholds of −0.5/day and −1.0/day; for UUI episodes/day we evaluated thresholds of −0.3/day and −0.5/day; and for voided volume we evaluated thresholds of +10, +20, and +30 mL.

2.6. Statistical Analysis

All analyses were performed using R version 4.5.0 (R Foundation for Statistical Computing, Vienna, Austria).

3. Results

3.1. Included Trials and Bayesian Cross‐Trial Synthesis of Pooled Effects, Heterogeneity, and Probabilities

We included three randomized, double‐blind, placebo‐controlled trials, comprising 1903 participants randomized to the arms included in this analysis; however, the number of analyzable patients varied by endpoint (Table S1). Figure 1 summarizes trial‐specific estimates and pooled effects at Week 12. Table 1 provides a concise summary of the pooled effects, posterior probabilities of benefit, and decision‐threshold exceedance probabilities.

FIGURE 1.

Trial‐specific and pooled effects of vibegron monotherapy at 12 weeks. Forest plots of placebo‐adjusted least‐squares (LS) mean differences in change from baseline. Blue bars indicate trial‐specific estimates with 95% confidence intervals; red bars indicate pooled mean effects from Bayesian random‐effects models with 95% credible intervals. Negative values indicate improvement for diary frequency endpoints, and positive values indicate improvement for voided volume. CI, confidence interval; CrI, credible interval; LS, least‐squares; RE, random effects; UUI, urgency urinary incontinence.

TABLE 1.

Bayesian summary of pooled 12‐week efficacy outcomes for vibegron monotherapy versus placebo.

Endpoint	Pooled effect vs. placebo, posterior median (95% CrI)	Posterior probability of benefit	Posterior probability of achieving decision threshold(s)
Micturitions/day	−0.76/day (−1.33 to −0.28)	99.5%	≤ −0.5/day: 90.3% ≤ −1.0/day: 14.2%
Urgency episodes/day	−0.69/day (−1.35 to −0.22)	99.3%	≤ −0.5/day: 82.8% ≤ −1.0/day: 12.7%
UUI episodes/day	−0.44/day (−0.93 to −0.03)	98.0%	≤ −0.3/day: 81.8% ≤ −0.5/day: 35.4%
Voided volume per micturition	+24.09 mL (+15.78 to +32.51)	> 99.9%	≥ +10 mL: 99.5% ≥ +20 mL: 88.4% ≥ +30 mL: 5.8%

Note: Values are posterior medians (95% credible intervals [CrIs]) from Bayesian normal–normal random‐effects models. Benefit was defined as μ < 0 for diary‐frequency endpoints and μ > 0 for voided volume. Decision thresholds were prespecified interpretive anchors for the pooled placebo‐adjusted mean effect and are not validated minimal important differences or patient‐level responder criteria.

Abbreviation: UUI, urgency urinary incontinence.

3.2. Micturitions/Day

The Bayesian random‐effects synthesis produced the pooled average effect parameter μ (posterior median, 95% credible intervals [CrI]) of −0.76/day (95% CrI, −1.33 to −0.28). The posterior median heterogeneity was τ = 0.27 (95% CrI, 0.018–0.87). The probability of benefit [Pr(μ < 0)] was 99.5%. The probability of exceeding decision thresholds was Pr(μ ≤ −0.5/day) = 90.3% and Pr(μ ≤ −1.0/day) = 14.2%. Numerical summaries of μ, the thresholds, and τ are shown in Figures 2, 3, 4.

FIGURE 2.

Posterior distributions of pooled mean effects under Bayesian random‐effects synthesis. Red dashed vertical lines indicate the posterior median and the 95% credible interval (CrI). The probability of benefit is computed as Pr(μ < 0) for diary frequency endpoints and Pr(μ > 0) for voided volume. UUI, urgency urinary incontinence.

FIGURE 3.

Posterior probability of exceeding decision thresholds. Bars indicate posterior probabilities derived from the Bayesian random‐effects model. UUI, urgency urinary incontinence.

FIGURE 4.

Posterior distributions of between‐trial heterogeneity under the Bayesian random‐effects synthesis. The unit of heterogeneity parameter (τ) corresponds to the scale of each endpoint (episodes/day or mL). CrI, credible interval.

3.3. Urgency Episodes/Day

The pooled average effect was μ = −0.69/day (95% CrI, −1.35 to −0.22; Figure 2), with τ = 0.27 (posterior median; 95% CrI, 0.013–0.91; Figure 4). The benefit probability was Pr(μ < 0) = 99.3%. The probabilities of exceeding −0.5/day and −1.0/day were 82.8% and 12.7%, respectively (Figure 3).

3.4. UUI Episodes/Day

The pooled average effect was μ = −0.44/day (95% CrI, −0.93 to −0.03; Figure 2), with τ = 0.23 (posterior median; 95% CrI, 0.018–0.79; Figure 4). The benefit probability was Pr(μ < 0) = 98.0%. The probabilities of exceeding −0.3/day and −0.5/day were 81.8% and 35.4%, respectively (Figure 3).

3.5. Voided Volume Per Micturition

The pooled average effect was μ = +24.1 mL (95% CrI, 15.8–32.5; Figure 2), with τ = 3.4 (posterior median; 95% CrI, 0.15–14.7; Figure 4). The benefit probability was Pr(μ > 0) > 99.9%. The probabilities of exceeding +10, +20, and +30 mL were 99.5%, 88.4%, and 5.8%, respectively (Figure 3).

3.6. Sensitivity Analysis

In the dose‐restricted analyses, the direction of the pooled mean effect remained favorable, and probability metrics were broadly stable (Figures S1–S6). Sensitivity analyses with alternative heterogeneity priors yielded consistent direction and similar magnitude of μ (Figure 5). More diffuse priors widened credible intervals but had only small effects on benefit probabilities and threshold‐exceedance probabilities. These results suggest that, even acknowledging that τ estimation is intrinsically uncertain with a small number of trials, the main inferences about the efficacy of vibegron are reasonably robust to plausible prior specifications.

FIGURE 5.

Sensitivity analysis for prior assumptions on heterogeneity. Sensitivity analysis of pooled mean effects (μ) under alternative prior distributions for between‐trial heterogeneity, including narrow, wide, and very wide half‐normal priors. Points indicate posterior medians, and horizontal bars represent 95% credible intervals (CrI). UUI, urgency urinary incontinence.

3.7. Interpretation of Probabilities

Figure 6 shows the posterior predictive distributions of the true treatment effect in a new trial setting (θ _new) for each endpoint. The predictive medians (95% posterior predictive intervals) at Week 12 were: micturitions/day −0.76 (−1.79 to 0.17); urgency episodes/day −0.67 (−1.83 to 0.24); UUI episodes/day −0.43 (−1.33 to 0.37); and voided volume +24.09 mL (9.02–39.32). Because posterior predictive intervals incorporate uncertainty in both μ and τ, they were wider than the corresponding CrIs for μ. Notably, the pooled effect estimates and posterior benefit probabilities remained stable across model specifications.

FIGURE 6.

Posterior predictive distributions for a new trial setting. Posterior predictive distributions of the treatment effect (θ _new) expected in a new trial setting, incorporating uncertainty in both the pooled mean effect (μ) and between‐trial heterogeneity (τ). Solid curves represent predictive densities; red dashed vertical lines indicate predictive medians. Predictive intervals are wider than pooled credible intervals because they account for uncertainty in both μ and τ. CrI, credible interval; PrI, posterior predictive interval; UUI, urgency urinary incontinence.

4. Discussion

Using published aggregate data from three 12‐week, randomized, double‐blind, placebo‐controlled Phase III trials of vibegron monotherapy conducted in Japan, Korea, and multinational settings, we derived Bayesian posterior distributions that directly express treatment efficacy in probabilistic terms. At Week 12, the pooled average effect favored vibegron across all four major bladder diary endpoints, with consistently high posterior probabilities of benefit. By integrating evidence across trials and presenting results as probabilities rather than solely as point estimates with intervals, this analysis provides a decision‐oriented interpretation of treatment effects.

International guidelines recommend β‐3 adrenoceptor agonists as a major pharmacological option for OAB, alongside antimuscarinics [9, 10]. Mirabegron, the first β‐3 agonist approved for OAB, is supported by multiple Phase III trials and long‐term evaluations demonstrating efficacy and tolerability [11, 12, 13]. Systematic reviews and network meta‐analyses have further supported the clinical utility of β‐3 agonists [14, 15]. Within this therapeutic landscape, our results indicate that vibegron has a high probability of benefit for key bladder diary outcomes and adds a probabilistic interpretation that extends beyond statistical significance.

Network meta‐analyses across drug classes provide mean differences, odds ratios, and ranking measures for efficacy and safety [14, 15, 16]. However, they typically emphasize point and interval estimates, and explicit reporting of the probability of exceeding clinically interpretable thresholds is uncommon. The present study did not aim to compare drugs; instead, it synthesized the vibegron evidence base and reported probabilities aligned with clinically interpretable questions, such as “the probability that the pooled average effect exceeds −0.5 episodes/day.” This probability‐based presentation can complement existing comparative meta‐analyses while clearly indicating that the probabilities pertain to uncertainty in the pooled average effect.

Interpreting continuous bladder diary endpoints also requires considering the magnitude of change that is clinically important. Frameworks for determining MID have been proposed [17], and meaningful‐change analyses in OAB have related diary improvements to patient‐reported anchors such as global impression or symptom‐bother measure [8, 18]. However, MID values are typically defined in terms of within‐group change from baseline and are not expressed as placebo‐adjusted differences. Accordingly, because no established MID thresholds are available for placebo‐adjusted contrasts, we specified decision thresholds that approximate the margin between the magnitude of change observed with placebo and the magnitude considered to represent clinically important improvement. The absolute decision thresholds used in this study are not presented as validated MID values; rather, they serve as transparent anchors to aid interpretation of the pooled, placebo‐adjusted mean effect. Importantly, an absolute change (episodes/day or mL) can have different clinical meaning depending on baseline symptom burden. In EMPOWUR, meaningful responder definitions were expressed as percentage reductions from baseline (e.g., micturitions ≥ 15% reduction [post hoc], urgency episodes ≥ 50% reduction, and UUI episodes ≥ 75% reduction) [8]. Because our estimand is a placebo‐adjusted mean difference in change from baseline, these responder definitions are not directly comparable to our thresholds, but they provide useful context for the magnitude of change that patients may perceive as meaningful.

For illustration, if baseline micturitions are approximately 8–10/day, a pooled mean difference of −0.5/day corresponds to only about a 5%–6% reduction, whereas −1.0/day corresponds to about a 10%–12% reduction. Similarly, if baseline UUI is about three episodes/day, −0.5/day corresponds to about a 17% reduction, which is substantially smaller than the ≥ 75% responder definition used in EMPOWUR [8]. Accordingly, our probability of exceeding absolute thresholds should be interpreted primarily as evidence about (i) the direction of the average placebo‐adjusted effect and (ii) whether that average effect surpasses modest, prespecified anchors, rather than as the probability that most patients achieve a percentage‐based “meaningful responder” status. The same principle applies to voided volume: increases of +10 to +20 mL are directionally consistent with improved bladder storage function, but the patient‐perceived meaningfulness of a given mL increase likely varies with baseline capacity and symptom bother and would ideally be interpreted alongside patient‐reported outcomes when such data are available [8, 17].

Methodologically, random‐effects meta‐analysis is a standard approach for synthesizing trial results in the presence of heterogeneity [19]. Bayesian extensions enable direct probability statements about pooled effects and a coherent framework for propagating uncertainty [20]. Bayesian evidence synthesis has also been advocated as being well aligned with decision‐relevant quantities in medical decision‐making [21]. In this synthesis, heterogeneity τ is particularly relevant because the included trials differed in vibegron dose (50 vs. 75 mg) and geographic setting (Japan, Korea, and multinational) and may also have differed in baseline symptom severity and placebo response, all of which can influence placebo‐adjusted diary changes. With only three trials, the data are insufficient to formally separate these sources of heterogeneity; τ should therefore be interpreted as summarizing net between‐trial variability rather than attributing it to any single factor. Given that only three trials were included (k = 3), estimation of τ is inherently imprecise and potentially sensitive to prior assumptions; this uncertainty should be considered when interpreting threshold‐exceedance probabilities. When the clinical question concerns the effect that might be expected in a specific practice setting, posterior predictive summaries based on θ _new, which incorporate τ, may be more directly informative than μ. Because the included trials differed by dose, exchangeability should be interpreted as comparability rather than strict identity. Dose‐restricted analyses offer a transparent means of demonstrating that overall conclusions are not driven by a single study and of illustrating the impact of combining different doses. As additional trials conducted in more homogeneous settings become available, future updates may enable more direct investigation of heterogeneity sources and potentially reduce predictive uncertainty.

This study has several limitations. First, the analysis was based on published aggregate data; therefore, patient‐level heterogeneity and individualized response prediction could not be assessed, and we could not directly link diary changes (e.g., −0.5/day or +20 mL) to patient‐reported satisfaction or quality‐of‐life improvements. Second, only three trials with differing doses and geographic regions were synthesized. Under the exchangeability assumption, μ should be interpreted as an across‐setting average effect. In small meta‐analyses, estimation of τ may be imprecise and sensitive to prior assumptions, potentially influencing predictive and threshold‐based interpretations. In addition, exploration of effect modification by race/ethnicity or baseline symptom severity was not feasible with aggregate‐level data. Third, SEs were back‐calculated from reported 95% CIs using a normal approximation; minor discrepancies may arise if the original analyses employed model‐based variance estimators. Finally, the thresholds predefined in this analysis were used as interpretative anchors rather than validated MID values, and they are not intended to reproduce the percentage‐based meaningful responder definitions used in EMPOWUR [8].

5. Conclusions

Bayesian cross‐trial synthesis of three placebo‐controlled vibegron monotherapies indicates a high probability that the pooled average effect is beneficial for key bladder diary endpoints at Week 12. Probability‐based summaries of pooled effects and threshold attainment can complement conventional reporting and provide clinically interpretable information for decision‐making in OAB pharmacotherapy.

Author Contributions

Takanobu Yamamoto: writing – original draft, conceptualization, methodology, software, formal analysis, data curation, visualization. Soichiro Yoshida: writing – review and editing, conceptualization, methodology, data curation, supervision, project administration. Shugo Yajima: writing – review and editing, methodology. Hiroshi Fukushima: writing – review and editing, methodology. Yosuke Yasuda: writing – review and editing. Yuma Waseda: writing – review and editing. Hiroyuki Sato: writing – review and editing. Akihiro Hirakawa: writing – review and editing. Hiroyuki Honda: writing – review and editing. Yoshinobu Komai: writing – review and Editing. Tomohiro Matsuo: writing – review and editing. Naoki Wada: writing – review and editing. Hitoshi Masuda: writing – review and editing, supervision, project administration. Yasuhisa Fujii: writing – review and editing, supervision, project administration.

Funding

The authors have nothing to report.

Ethics Statement

This study was based on previously published aggregate data and did not involve new human participants or identifiable patient information. Therefore, institutional review board approval was not required.

Consent

The authors have nothing to report.

Conflicts of Interest

Akihiro Hirakawa has received consulting fees from Kissei Pharmaceutical Co. Ltd. outside the submitted work. Tomohiro Matsuo and Naoki Wada received speaker honoraria from Kissei Pharmaceutical Co. Ltd. Tomohiro Matsuo, Naoki Wada, and Hitoshi Masuda received speaker honoraria from KYORIN Pharmaceutical Co. Ltd. Soichiro Yoshida has received research funding from Kissei Pharmaceutical Co. Ltd. and KYORIN Pharmaceutical Co. Ltd. The other authors declare no conflicts of interest.

Supporting information

Table S1: 12‐Week vibegron monotherapy (placebo‐adjusted) continuous endpoints.

Table S2: Computational settings used in the R script.

Figure S1: Trial‐specific and pooled effects of vibegron 50 mg monotherapy at 12 weeks.

Figure S2: Posterior distributions of pooled mean effects under Bayesian random‐effects synthesis.

Figure S3: Posterior probability of exceeding decision thresholds.

Figure S4: Posterior distributions of between‐trial heterogeneity under the Bayesian random‐effects synthesis.

Figure S5: Sensitivity analysis for prior assumptions on heterogeneity.

Figure S6: Posterior predictive distributions for a new trial setting.

Data Availability Statement

The data supporting the findings of this study were derived from previously published studies and are summarized within the manuscript and Supporting Information. No new individual‐level data were generated.