Comparative Efficacy of Aldosterone‐Related Sodium‐Retention Pathway Therapies for Resistant Hypertension: A Drug‐Level Network Meta‐Analysis of Randomized Controlled Trials

ABSTRACT

Mineralocorticoid receptor antagonists (MRAs), aldosterone synthase inhibitors (ASIs), and epithelial sodium channel (ENaC) blockade target different levels of the aldosterone‐related sodium‐retention pathway in resistant hypertension (RH), but the comparative efficacy of individual agents remains uncertain. We conducted a focused drug‐level network meta‐analysis of randomized controlled trials (RCTs) comparing pathway‐directed therapies in adults with RH. PubMed, Scopus, Embase, Cochrane CENTRAL, Cochrane Reviews, and ClinicalTrials.gov were searched from inception to July 20, 2026. The primary outcome was change in office systolic blood pressure (BP). Secondary outcomes included office diastolic BP and 24‐hour ambulatory systolic and diastolic BP. Ten RCTs including 2865 participants were analyzed. In the primary drug‐level analysis, amiloride, spironolactone, baxdrostat, eplerenone, and lorundrostat significantly reduced office systolic BP compared with placebo, with mean differences (MDs) of −14.08, −10.78, −9.09, −8.14, and −6.80 mmHg, respectively. Osilodrostat showed a statistically uncertain effect, with an MD of −2.61 mmHg. At the class level, both MRAs and ASIs showed clinically meaningful reductions in office systolic BP. Secondary outcomes were generally consistent with the primary analysis, although fewer trials contributed to these networks. Sensitivity analyses did not materially change the primary findings. These results support the aldosterone‐related sodium‐retention pathway as an important therapeutic target in RH. MRAs remain the established reference add‐on therapy, whereas newer ASIs and amiloride may represent pathway‐based alternatives for selected patients.

Trial Registration: INPLASY202670099

1. Introduction

Resistant hypertension (RH) is defined as blood pressure (BP) remaining above target despite an optimized three‐drug antihypertensive regimen that includes a diuretic, or as BP controlled only with the use of four or more antihypertensive agents [1]. In a large meta‐analysis, the estimated prevalence of true RH was approximately 10.3% among treated hypertensive patients, highlighting its substantial clinical burden [2]. Current major hypertension guidelines consistently position mineralocorticoid receptor antagonists (MRAs), particularly spironolactone, as the preferred fourth‐line pharmacological therapy for RH after optimization of standard triple therapy with renin–angiotensin system blockade, calcium channel blockade, and diuretic therapy. Eplerenone and amiloride are generally considered alternative options, although their positioning and recommendations regarding renal function and serum potassium vary across guidelines (Table 1) [1, 3, 4]. The established role of spironolactone is supported by the PATHWAY‐2 trial, which demonstrated substantial BP reductions in patients with RH, and by subsequent mechanistic analyses linking this benefit to reversal of aldosterone‐related sodium retention [5, 6].

TABLE 1.

Current guideline positioning of aldosterone‐related sodium‐retention pathway therapies in resistant hypertension.

Guideline	Spironolactone	Eplerenone	Amiloride	Key restrictions/precautions	Newer ASIs
2023 ESH [4]	Preferred fourth‐line; 25–50 mg/day	Alternative; dose not specified	Alternative; 10–20 mg/day	Caution with eGFR <45 mL/min/1.73 m2 or K+ >4.5 mmol/L; MRA contraindicated with eGFR <30 mL/min/1.73 m2	Baxdrostat discussed as an emerging option; not established for routine treatment
2024 ESC [3]	Preferred fourth‐line; 25–50 mg/day	Alternative; 50–200 mg/day, often BID	Later‐line alternative; dose not specified	Spironolactone: eGFR ≥30 mL/min/1.73 m2 and K+ ≤4.5 mmol/L; steroidal MRA contraindicated with eGFR <30 mL/min/1.73 m2	Baxdrostat and lorundrostat discussed as investigational therapies; not established for routine treatment
2025 AHA/ACC [1]	Preferred fourth‐line; 25–50 mg/day	Alternative; 25–50 mg BID	Alternative if MRA unsuitable; 10–20 mg/day	MRA recommendation applies to eGFR ≥45 mL/min/1.73 m2	Not established for routine treatment

Abbreviations: AHA/ACC: American Heart Association/American College of Cardiology; ASI: aldosterone synthase inhibitor; BID: twice daily; eGFR: estimated glomerular filtration rate; ESC: European Society of Cardiology; ESH: European Society of Hypertension; K⁺: serum potassium; MRA: mineralocorticoid receptor antagonist; RH: resistant hypertension.

Aldosterone‐related sodium retention can be targeted at several points along the same biological pathway. In addition to mineralocorticoid receptor blockade, upstream inhibition of aldosterone synthesis with aldosterone synthase inhibitors (ASIs) may reduce aldosterone production, whereas downstream epithelial sodium channel (ENaC) blockade with amiloride may directly inhibit distal tubular sodium reabsorption. In recent years, the development of newer selective ASIs, including baxdrostat and lorundrostat, has expanded the therapeutic landscape of aldosterone‐targeted treatment for hypertension [7, 8]. Amiloride and ASIs have both demonstrated BP‐lowering effects in patients with resistant or uncontrolled hypertension [6, 7, 8]. A recent network meta‐analysis (NMA) by Xiao et al. evaluated a broad range of pharmacological and interventional treatments for RH and found that both ASIs and MRAs were associated with significant BP reductions [9]. Nevertheless, treatment nodes were defined at the class level, leaving the comparative efficacy of individual agents within the broader aldosterone‐related sodium‐retention pathway uncertain.

Therefore, we conducted a focused drug‐level NMA of randomized controlled trials (RCTs) to compare pharmacological therapies targeting the aldosterone‐related sodium‐retention pathway in RH. By evaluating MRAs, ASIs, and amiloride‐mediated ENaC blockade within a unified comparative framework, this study aimed to compare the short‐term BP‐lowering efficacy of individual agents and address a clinically relevant evidence gap not fully resolved by previous studies.

2. Methods

This NMA was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta‐Analyses extension statement for network meta‐analyses (PRISMA‐NMA) [10]. The protocol was registered in the International Platform of Registered Systematic Review and Meta‐Analysis Protocols (INPLASY; registration number: INPLASY202670099) [11]. Because this study synthesized aggregate data from previously published trials, institutional review board approval and informed consent were not required.

2.1. Literature Search and Study Selection

A systematic literature search was conducted in PubMed, Scopus, Cochrane CENTRAL, Cochrane Reviews, Embase, and ClinicalTrials.gov from database inception to July 20, 2026, without language restrictions. The search was independently performed by two reviewers (YLT and YHW) and aimed to identify RCTs evaluating aldosterone‐related sodium‐retention pathway pharmacotherapies in patients with RH.

The search strategy combined terms related to RH, randomized trial design, and eligible pharmacological interventions, including MRAs, ASIs, and ENaC blockade. The complete search strategies for each database and registry are provided in Table S1. Two reviewers (YLT and YHW) independently screened titles and abstracts, followed by full‐text review of potentially eligible studies. Reference lists of relevant reviews and included articles were also examined to identify additional trials. Disagreements were resolved by discussion, with consultation of a third reviewer (YSL) when needed.

2.2. Eligibility Criteria

RCTs were eligible if they met the predefined PICO (population, intervention, comparison, outcome) framework. The population of interest was adults with RH, defined as uncontrolled BP despite treatment with at least three antihypertensive agents, including a diuretic. For trials enrolling broader hypertensive populations, only RH subgroup data were extracted. Studies were excluded if RH‐specific data could not be separately obtained.

Eligible interventions were pharmacological agents targeting the aldosterone‐related sodium‐retention pathway, including MRAs, ASIs, and ENaC blockade. The predefined eligible drugs were spironolactone, eplerenone, baxdrostat, lorundrostat, osilodrostat, and amiloride. Eligible comparators included placebo or another eligible active treatment within the predefined network.

Studies were excluded if they were non‐randomized, observational, single‐arm, or non‐original reports. Trials restricted to special RH populations, including chronic kidney disease, dialysis‐dependent kidney failure, obstructive sleep apnea syndrome, heart failure, or rheumatoid arthritis, were also excluded to reduce clinical heterogeneity and enhance transitivity.

2.3. Data Extraction and Outcome Definitions

Two reviewers (YLT and YHW) independently extracted data using a standardized form. Extracted information included study design, sample size, RH definition, participant characteristics, intervention and comparator details, dose, follow‐up duration, BP measurement method, and outcome data.

For continuous outcomes, mean changes from baseline and corresponding standard deviations, standard errors, or confidence intervals (CIs) were extracted when available. When multiple follow‐up time points were reported, data from the primary efficacy time point of each trial were used. For trials enrolling mixed hypertensive populations, only RH subgroup data were extracted. When multiple dose arms of the same drug were reported within a trial, they were combined into a single drug‐level node before analysis to avoid double‐counting of shared comparators. Discrepancies were resolved by discussion, with adjudication by a third reviewer (YSL) when necessary.

The primary outcome was change in office systolic BP from baseline. Secondary outcomes were changes in office diastolic BP, 24‐hour ambulatory systolic BP, and 24‐hour ambulatory diastolic BP. Office and ambulatory BP outcomes were analyzed separately because they represent different measurement contexts and were not considered interchangeable.

2.4. Risk of Bias Assessment

The risk of bias of included RCTs was independently assessed by two reviewers (YLT and YHW) using the Cochrane Risk of Bias 2 tool [12]. The evaluated domains included the randomization process, intervention adherence, missing outcome data, outcome measurement, and selective reporting. Overall risk of bias was categorized as low risk, some concerns, or high risk. Disagreements were resolved by consensus or by consultation with a third reviewer (CLC).

2.5. Treatment Nodes

The primary analysis used drug‐level treatment nodes, comprising placebo, spironolactone, eplerenone, baxdrostat, lorundrostat, osilodrostat, and amiloride. Mechanism‐based class‐level analyses were additionally performed for the primary outcome. Spironolactone and eplerenone were grouped as MRAs, whereas baxdrostat, lorundrostat, and osilodrostat were grouped as ASIs. Amiloride was excluded from the MRA‐versus‐ASI class‐level analysis because its mechanism differs from mineralocorticoid receptor blockade. An additional analysis excluding osilodrostat was performed to evaluate whether the ASI estimate was influenced by inclusion of this older agent.

2.6. Statistical Analysis

NMA was performed within a frequentist framework using MetaInsight (version 7.1.1, Complex Reviews Support Unit, National Institute for Health Research, London, UK), which implements the netmeta package in R software [13]. A random‐effects model was used as the primary model because clinical and methodological heterogeneity was expected across trials. Treatment effects for continuous BP outcomes were expressed as mean differences (MDs) with 95% CIs.

Network plots were generated to show the geometry of available evidence. For the primary outcome, a drug‐level forest plot and league table were generated, followed by a class‐level forest plot comparing MRAs, ASIs, and placebo. For secondary outcomes, forest plots were generated using the drug‐level node structure.

2.7. Assessment of Inconsistency and Transitivity

Statistical inconsistency was assessed by comparing direct and indirect evidence when closed loops were available. Inconsistency testing was performed only when the network structure allowed formal estimation. When networks were sparse or lacked evaluable closed loops, inconsistency tests were considered not estimable.

To appraise the plausibility of transitivity, we examined the distribution of available potential effect modifiers across treatment comparisons, including RH definition, baseline BP, intervention dose, follow‐up duration, and BP measurement method.

2.8. Sensitivity Analyses

Sensitivity analyses were conducted for the primary outcome of office systolic BP change. First, a fixed‐effect model was applied to assess whether the results were influenced by model specification. Second, for studies in which change‐score standard deviations were imputed, the analysis was repeated using a lower pre–post correlation coefficient of 0.5 instead of 0.8 to assess the influence of this assumption. Third, a study design‐related sensitivity analysis was performed by excluding crossover trials. Sensitivity analyses were performed only when the evidence network remained connected.

3. Results

3.1. Study Selection

The study selection process is summarized in the PRISMA flow diagram in Figure 1. The PRISMA‐NMA checklist is provided in Table S2, and the numbers of records retrieved from each database and registry are detailed in Table S1. After removal of duplicates and screening of titles and abstracts, 38 full‐text articles were assessed for eligibility, of which 28 were excluded. Ultimately, 10 RCTs involving 2865 participants were included in the qualitative synthesis and NMA [5, 8, 14, 15, 16, 17, 18, 19, 20, 21]. Full‐text excluded studies with reasons for exclusion are listed in Table S3 [6, 7, 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].

FIGURE 1.

PRISMA flow diagram of study selection. RCT: randomized controlled trial; RH: resistant hypertension.

3.2. Study Characteristics and Risk of Bias

The baseline characteristics of the included trials are summarized in Table 2. The included studies evaluated MRAs, ASIs, or amiloride as add‐on therapies, with follow‐up durations ranging from 6 to 26 weeks. Across treatment arms with available data, mean age ranged from 53 to 63.9 years, the proportion of male participants ranged from 51.1% to 84.6%, and mean body mass index ranged from 28.1 to 33.3 kg/m². Baseline office systolic BP ranged from 139 to 166 mmHg, and baseline 24‐hour ambulatory systolic BP, when reported, ranged from approximately 140.5 to 147.6 mmHg.

TABLE 2.

Baseline characteristics of included randomized controlled trials.

First author & year	Study name	Interventions	Participants in nodes	Mean age	Male	BMI (kg/m2)	Follow‐up duration	Baseline office BP (mmHg)	Baseline 24‐h BP (mmHg)
Karns 2013	NA	Osilodrostat (LCI699) Eplerenone Placebo	89 33 33	55.4 56.2 59.8	62.9% 57.6% 66.7%	33.3 33.2 31.9	8 weeks	152.4/90.0 153.8/89.1 153.4/90.1	NA
Oxlund 2013	NA	Spironolactone Placebo	61 58	62.9 63.9	75% 78%	32.0 31.5	16 weeks	144/79 139/76	144/78 143/78
Václavík 2014	ASPIRANT‐EXT	Spironolactone Placebo	74 76	60.4 59.7	67.6% 63.2%	32.2 32.3	8 weeks	154.9/93.1 153.3/91.1	144.5/83.2 141.2/81.3
Williams 2015	PATHWAY‐2	Spironolactone Placebo	285 274	61.4	69%	NA	12 weeks	157/90	147.6/84.2
Kalizki 2017	NA	Eplerenone Placebo	25 26	62.2 57.7	76% 84.6%	29.3 28.1	26 weeks	166/91 159/94	143/82 143/86
Freeman 2023	BrigHTN	Baxdrostat Placebo	205 69	61.8 63.8	53.9% 61%	32.8 32.1	12 weeks	147.5/87.8 148.9/88.2	NA
Lee 2025	NA	Amiloride Spironolactone	56 58	53 55	70.7% 70%	30.3 29.0	12 weeks	145.8/87.6 144.5/88.1	141.5/86.1 142.3/87.0
Flack 2025	BaxHTN	Baxdrostat Placebo	386 192	60.8 61.9	62.6% 61.4%	31.3 31.1	12 weeks	149.4/86.9 149.0/85.8	NA
Saxena 2025	Launch‐HTN	Lorundrostat Placebo	490 158	61.6 61.8	53.8% 51.1%	32.9 32.6	6 weeks	148.6/NA 149.0/NA	NA
Azizi 2026	Bax24	Baxdrostat Placebo	108 109	60 61	65% 64%	32.4 31.6	12 weeks	146.9/85.4 148.3/85.7	140.5/79.7 141.8/80.0

Abbreviations: 24‐h: 24‐hour ambulatory; BMI: body mass index; BP: blood pressure; NA: not available.

Risk of bias was assessed for all included trials. Six trials were judged to be at low risk of bias, four had some concerns, and none was judged to be at high risk of bias. Figure S1 presents the summary risk‐of‐bias assessment, and detailed domain‐level judgments are provided in Table S4.

3.3. Network Geometry

The evidence network for the primary outcome of office systolic BP included 10 RCTs and 7 treatment nodes: placebo, spironolactone, eplerenone, baxdrostat, lorundrostat, osilodrostat, and amiloride. Placebo served as the central comparator.

Most active treatments were connected to placebo through direct comparisons, whereas amiloride was connected to the network through its comparison with spironolactone. Figure 2 illustrates the network geometry for office systolic BP.

FIGURE 2.

Network plot for office systolic blood pressure. Each node represents a treatment, and each line represents a direct comparison between treatments. Node size is proportional to the number of participants, and line thickness is proportional to the number of direct comparisons.

In the mechanism‐based class‐level analysis for office systolic BP, treatments were grouped into three nodes: placebo, MRAs, and ASIs. This network included 9 trials and was used to evaluate the consistency of class‐level estimates with the primary drug‐level analysis. Figure S2 presents the class‐level network.

For secondary outcomes, the office diastolic BP network included 6 trials and 5 treatment nodes, as shown in Figure S3. The 24‐hour ambulatory systolic and diastolic BP outcomes shared the same network structure, comprising 5 trials and 6 treatment nodes. Figure S4 shows the 24‐hour ambulatory BP network. These secondary networks were smaller than the primary office systolic BP network and were largely placebo‐centered.

3.4. Primary Outcome: Office Systolic Blood Pressure

In the primary drug‐level NMA, most active treatments significantly reduced office systolic BP compared with placebo. Figure 3 presents the forest plot for the primary analysis. The MDs versus placebo were −14.08 mmHg for amiloride (95% CI, −17.40 to −10.77), −10.78 mmHg for spironolactone (95% CI, −13.14 to −8.42), −9.09 mmHg for baxdrostat (95% CI, −11.29 to −6.88), −8.14 mmHg for eplerenone (95% CI, −14.56 to −1.72), and −6.80 mmHg for lorundrostat (95% CI, −8.47 to −5.13). Osilodrostat showed a statistically uncertain effect, with an MD of −2.61 mmHg versus placebo (95% CI, −8.75 to 3.52). The league table summarizing direct and indirect comparisons between all treatment nodes is presented in Table 3.

FIGURE 3.

Forest plot of drug‐level network meta‐analysis for office systolic blood pressure. Results are expressed as mean differences with 95% confidence intervals. Negative values indicate greater blood pressure reduction compared with placebo. CI: confidence interval; MD: mean difference.

TABLE 3.

League table of drug‐level network meta‐analysis for office systolic blood pressure.

Amiloride	−3.30 [−5.63, −0.97]	−	−	−	−	−
−3.30 [−5.63, −0.97]	Spironolactone	−	−	−	−	−10.78 [−13.14, −8.42]
−5.00 [−8.98, −1.02]	−1.70 [−4.93, 1.53]	Baxdrostat	−	−	−	−9.09 [−11.29, −6.88]
−5.95 [−13.17, 1.28]	−2.65 [−9.49, 4.20]	−0.95 [−7.74, 5.84]	Eplerenone	−	−6.42 [−13.08, 0.24]	−8.14 [−14.56, −1.72]
−7.28 [−10.99, −3.57]	−3.98 [−6.88, −1.09]	−2.29 [−5.05, 0.48]	−1.34 [−7.98, 5.30]	Lorundrostat	−	−6.80 [−8.47, −5.13]
−11.47 [−18.45, −4.49]	−8.17 [−14.75, −1.59]	−6.47 [−13.00, 0.05]	−5.52 [−11.72, 0.67]	−4.19 [−10.55, 2.18]	Osilodrostat	−3.48 [−10.06, 3.10]
−14.08 [−17.40, −10.77]	−10.78 [−13.14, −8.42]	−9.09 [−11.29, −6.88]	−8.14 [−14.56, −1.72]	−6.80 [−8.47, −5.13]	−2.61 [−8.75, 3.52]	Placebo

Note: Values are mean differences in office systolic blood pressure change from baseline with 95% confidence intervals, expressed in mmHg. Comparisons should be read as the column‐defining treatment versus the row‐defining treatment. Negative values indicate greater blood pressure reduction with the column‐defining treatment. Estimates from pairwise meta‐analyses are presented above the diagonal line, and estimates from network meta‐analyses are presented below the diagonal line.

3.5. Class‐Level Analysis of Office Systolic Blood Pressure

In the class‐level analysis of office systolic BP, both MRAs and ASIs were associated with significant reductions compared with placebo. The MDs were −10.88 mmHg for MRAs (95% CI, −13.29 to −8.48) and −7.56 mmHg for ASIs (95% CI, −9.24 to −5.89). Figure 4 presents the forest plot for the class‐level analysis. In an additional analysis excluding osilodrostat, the estimate for ASIs remained similar (MD, −7.82 mmHg; 95% CI, −9.39 to −6.25), suggesting that the class‐level ASI estimate was not materially influenced by osilodrostat. This analysis is presented in Figure S5.

FIGURE 4.

Forest plot of class‐level network meta‐analysis for office systolic blood pressure. Mineralocorticoid receptor antagonists included spironolactone and eplerenone, whereas aldosterone synthase inhibitors included baxdrostat, lorundrostat, and osilodrostat. Results are expressed as mean differences with 95% confidence intervals. Negative values indicate greater blood pressure reduction compared with placebo. ASI: aldosterone synthase inhibitor; CI: confidence interval; MD: mean difference; MRA: mineralocorticoid receptor antagonist.

3.6. Secondary Outcomes

For office diastolic BP, baxdrostat and spironolactone were associated with significant reductions compared with placebo. The MDs were −5.00 mmHg for baxdrostat (95% CI, −7.66 to −2.34) and −4.14 mmHg for spironolactone (95% CI, −5.55 to −2.72). The estimates for eplerenone (MD, −2.49 mmHg; 95% CI, −5.92 to 0.94) and osilodrostat (MD, −0.50 mmHg; 95% CI, −3.99 to 2.99) were statistically uncertain. Figure S6 presents the forest plot for the office diastolic BP analysis.

For 24‐hour ambulatory systolic BP, eplerenone, baxdrostat, spironolactone, and amiloride were associated with significant reductions compared with placebo. The MDs were −14.70 mmHg for eplerenone (95% CI, −21.88 to −7.52), −14.00 mmHg for baxdrostat (95% CI, −17.07 to −10.93), −9.54 mmHg for spironolactone (95% CI, −12.16 to −6.93), and −8.44 mmHg for amiloride (95% CI, −12.91 to −3.98). Osilodrostat showed a statistically uncertain effect (MD, −4.44 mmHg; 95% CI, −10.46 to 1.58). The corresponding forest plot is shown in Figure S7.

For 24‐hour ambulatory diastolic BP, eplerenone, baxdrostat, amiloride, and spironolactone were associated with significant reductions compared with placebo. The MDs were −9.40 mmHg for eplerenone (95% CI, −14.23 to −4.57), −6.80 mmHg for baxdrostat (95% CI, −8.75 to −4.85), −3.79 mmHg for amiloride (95% CI, −6.84 to −0.74), and −3.69 mmHg for spironolactone (95% CI, −5.48 to −1.90). The estimate for osilodrostat was statistically uncertain (MD, −1.72 mmHg; 95% CI, −5.79 to 2.35). Figure S8 shows the forest plot for the 24‐hour ambulatory diastolic BP analysis.

3.7. Inconsistency Assessment

Formal inconsistency testing was limited by the sparse network structure. No statistically significant inconsistency was detected where testing was estimable, including the primary office systolic BP network, the office diastolic BP network, and the class‐level office systolic BP analysis. Inconsistency testing was not estimable for the 24‐hour ambulatory systolic or diastolic BP networks because evaluable closed loops were absent.

3.8. Sensitivity Analyses

Sensitivity analyses supported the robustness of the primary office systolic BP findings. The fixed‐effect model produced estimates identical to those from the primary random‐effects model. Alternative assumptions for imputed change‐score standard deviations did not alter the point estimates, direction of effect, or statistical significance of the primary findings. Excluding crossover trials also did not materially change the primary results. Figure S9 shows the forest plot for the analysis excluding crossover trials.

4. Discussion

4.1. Principal Findings

In this focused drug‐level NMA of 10 RCTs including 2865 patients with RH, most active therapies targeting the aldosterone‐related sodium‐retention pathway significantly reduced office systolic BP compared with placebo. At the drug level, significant reductions were observed with amiloride, spironolactone, baxdrostat, eplerenone, and lorundrostat, whereas the estimate for osilodrostat was statistically uncertain. In the class‐level analysis, both MRAs and ASIs showed clinically meaningful reductions in office systolic BP. Secondary outcomes were generally consistent with the primary analysis, although fewer trials contributed to office diastolic and 24‐hour ambulatory BP outcomes. Sensitivity analyses did not materially change the primary findings. Overall, our results support MRAs as established add‐on therapies for RH and suggest that ASIs, particularly baxdrostat, may provide meaningful short‐term BP reductions.

4.2. Comparison With Prior Studies

Previous NMAs have examined RH treatment from different perspectives. Tian et al. performed a drug‐level NMA of pharmacological and interventional treatments and reported favorable BP‐lowering effects with spironolactone, although newer ASIs and amiloride were not included [47]. Xiao et al. subsequently evaluated multiple treatment classes, including MRAs and ASIs, but their analysis was primarily class based rather than drug based [9]. ASI‐focused NMAs have also compared different ASI regimens and suggested dose‐related BP reductions, but they did not directly compare these agents with MRAs or amiloride within an RH‐specific network [48]. Our NMA therefore provides a complementary drug‐level analysis focused on aldosterone‐related sodium‐retention pathway pharmacotherapies in patients with RH.

4.3. Mineralocorticoid Receptor Antagonists

Our findings are consistent with the established role of MRAs as add‐on therapy for RH after optimization of standard triple antihypertensive treatment [49]. This effect is biologically plausible because sodium retention, volume expansion, and inappropriate aldosterone activity are important contributors to RH [50]. However, conventional steroidal MRAs are limited by adverse effects such as hyperkalemia, menstrual irregularities, and gynecomastia, particularly in patients with chronic kidney disease [31, 51]. Although newer non‐steroidal MRAs have shown BP‐lowering potential, they were not included in the present NMA because no eligible RCTs met our predefined inclusion criteria [35, 52, 53]. These limitations support the need for alternative therapies for patients who are intolerant of or unsuitable for MRA therapy.

4.4. Aldosterone Synthase Inhibitors

ASIs represent an upstream approach to targeting the aldosterone‐related sodium‐retention pathway. Unlike MRAs, which block receptor activation, ASIs reduce aldosterone production and may therefore be particularly relevant in patients with RH characterized by inappropriate aldosterone excess [54]. In this setting, suppressing aldosterone synthesis may provide a mechanistically targeted alternative to receptor blockade.

The interpretation of ASI efficacy requires some caution. A recent NMA of ASIs reported dose‐dependent BP reductions across baxdrostat, lorundrostat, and osilodrostat [48]. In our analysis, different doses of the same ASI were combined into single drug‐level nodes; therefore, the estimates should be interpreted as overall drug‐level effects rather than dose‐specific effects. Higher‐dose ASI regimens may achieve greater BP reductions than reflected by pooled drug‐level estimates, but this potential gain must be balanced against safety concerns, particularly hyperkalemia [55]. Although newer ASIs may provide meaningful short‐term BP reductions, equivalence with MRAs cannot be inferred from the current evidence.

4.5. Amiloride and Epithelial Sodium Channel Blockade

Amiloride provides a downstream approach to targeting sodium retention by directly blocking ENaC in the distal nephron. Mechanistic data from PATHWAY‐2 support sodium retention as an important driver of RH, providing a biological rationale for ENaC blockade in this setting [6]. This mechanism may be particularly relevant in patients with a Liddle‐like phenotype, characterized by low renin and low aldosterone levels, in whom ENaC‐mediated sodium reabsorption may contribute to sodium retention despite the absence of overt aldosterone excess [56].

This downstream mechanism also distinguishes amiloride from both MRAs and ASIs. Because amiloride acts at the level of ENaC, its BP‐lowering effect may be less dependent on upstream aldosterone activity. However, the evidence base for amiloride in RH remains limited.

4.6. Clinical Implications

The main clinical implication of this analysis is that add‐on therapy for RH may be better understood through a pathway‐based framework rather than a purely class‐based approach. Aldosterone synthesis, mineralocorticoid receptor activation, and downstream ENaC‐mediated sodium reabsorption represent related but distinct therapeutic targets. This framework may be useful when selecting add‐on therapy for patients with persistent hypertension despite optimized standard treatment or for those who cannot tolerate conventional MRA therapy.

Future trials should move beyond broad treatment comparisons and incorporate phenotype‐guided designs. Biomarkers such as renin, aldosterone, potassium, kidney function, and indices of sodium retention may help identify patients who are more likely to benefit from MRAs, ASIs, or amiloride‐mediated ENaC blockade. Such an approach may clarify the optimal role of each pathway‐directed therapy in the management of RH.

4.7. Limitations

This study has several limitations. First, the evidence network was sparse and largely placebo‐centered, with few direct head‐to‐head comparisons between active treatments. Therefore, several estimates relied on indirect evidence. To reduce this concern, we used a focused eligibility framework, predefined drug‐level treatment nodes, and assessed inconsistency whenever the network structure allowed formal estimation. Second, some treatment nodes were supported by single or few trials, resulting in less precise estimates. We therefore interpreted the findings as comparative evidence for short‐term BP reduction rather than definitive evidence of treatment superiority. Sensitivity analyses were also performed for the primary outcome and did not materially change the main findings. Third, the included trials mainly assessed short‐term BP outcomes, and long‐term cardiovascular or renal outcomes could not be evaluated. Finally, safety outcomes were not quantitatively compared because adverse event reporting was incomplete and inconsistent across trials. This is clinically relevant because hyperkalemia and renal function changes are important considerations when using therapies targeting the aldosterone‐related sodium‐retention pathway.

4.8. Future Perspectives

Although NMA enables comparisons among therapies that have not been directly compared in randomized trials, indirect evidence cannot substitute for adequately powered head‐to‐head studies. Future trials of newer pathway‐directed therapies should therefore move beyond placebo‐controlled efficacy studies and directly compare these agents with established treatments, particularly spironolactone. Further research should also clarify optimal dosing and identify patient subgroups most likely to benefit from ASIs or amiloride. Longer‐term studies evaluating safety, tolerability, and cardiovascular and renal outcomes, in addition to BP reduction, are also needed to better define their clinical positioning and inform treatment selection in RH.

5. Conclusions

In this focused drug‐level NMA of RCTs in RH, therapies targeting the aldosterone‐related sodium‐retention pathway were associated with clinically meaningful short‐term BP reductions, supporting this pathway as an important therapeutic target. MRAs remain the established add‐on therapy, whereas newer ASIs and amiloride may represent pathway‐based alternatives for selected patients, although further head‐to‐head trials are needed to better define their comparative clinical roles.

Author Contributions

Yi‐Liang Tsou: conceptualization, literature search, study selection, data extraction, risk of bias assessment, statistical analysis, data interpretation, writing – original draft preparation, supervision. Yu‐Hung Wang: literature search, study selection, data extraction, risk of bias assessment, data interpretation, writing – review and editing. Yi‐Siou Lin: data interpretation, statistical analysis, writing – review and editing. Chia‐Liang Chen: data interpretation, writing – review and editing. All authors read and approved the final manuscript.

Funding

This research received no specific grant from any funding agency in the public, commercial, or not‐for‐profit sectors.

Ethics Statement

Institutional review board approval was not required because this study synthesized aggregate data from previously published studies.

Consent

Patient consent was not required because no individual patient‐level data were used.

Conflicts of Interest

The authors declare no conflicts of interest.

Permission to Reproduce Material from Other Sources

No copyrighted material from other sources was reproduced in this manuscript.

Supporting information

Supporting Information: jch70365‐sup‐0001‐SuppMat.docx

Data Availability Statement

The data used in this study were extracted from published articles and publicly available sources and are available from the corresponding author upon reasonable request.