Comparative efficacy and pharmacological heterogeneity of individual potassium-competitive acid blockers: A systematic review and network meta-analysis

Figure 1 Preferred Reporting Items for Systematic reviews and Meta-Analyses flow diagram illustrating the study selection process. The systematic literature search identified 2441 records from four databases (PubMed: n = 111, EMBASE: n = 1107, Cochrane Library: n = 703, Web of Science: n = 520). After removing 1291 duplicates, 1150 unique records were screened by title and abstract. Following full-text review of 445 articles, 405 were excluded (narrative reviews: n = 6, incomplete data: n = 231, study protocols: n = 11, systematic reviews/guidelines: n = 157). Finally, 40 randomized controlled trials were included: (1) 30 studies for Helicobacter pylori eradication; and (2) 10 studies for erosive esophagitis healing.

Figure 2 Forest plot of network meta-analysis comparing potassium-competitive acid blockers vs proton pump inhibitors for Helicobacter pylori eradication. Each row represents a randomized controlled trial, with squares indicating point estimates of odds ratio and horizontal lines representing 95%CIs. Diamond at the bottom shows the pooled effect estimate. Overall heterogeneity was low (I² = 15.16%). P-CAB: Potassium-competitive acid blocker; PPI: Proton pump inhibitor.

Figure 3 Forest plot of individual potassium-competitive acid blockers (vonoprazan, tegoprazan, keverprazan) vs proton pump inhibitor for clarithromycin-resistant Helicobacter pylori eradication. Each row represents a randomized controlled trial, with squares indicating point estimates of odds ratio (OR) and horizontal lines representing 95%CIs. Diamond at the bottom shows the pooled effect estimate. Five studies reported clarithromycin resistance subgroup data. High-acid dissociation constant; potassium-competitive acid blockers (vonoprazan, keverprazan) and low-acid dissociation constant tegoprazan achieved pooled OR = 3.32 (95%CI: 1.79-6.16) compared with proton pump inhibitors (PPIs). Vonoprazan-based regimens (3 arms) achieved 65.8%-82.0% eradication in clarithromycin-resistant infections vs 31.9%-40.0% with PPIs. Keverprazan showed 83.5% vs 77.0% with a not significant OR of 1.51 (95%CI: 0.83-2.76), likely attenuated by the high baseline eradication rate in the Chinese PPI arm. Tegoprazan showed 47.8% vs 35.5% (OR = 1.67, not significant). Heterogeneity was substantial (I² = 72.3%, P = 0.006), reflecting geographic and methodological differences. P-CAB: Potassium-competitive acid blocker; PPI: Proton pump inhibitor.

Figure 4 Forest plot of network meta-analysis comparing potassium-competitive acid blockers vs proton pump inhibitors for 8-week erosive esophagitis healing. Each row represents a randomized controlled trial, with squares indicating point estimates of odds ratio and horizontal lines representing 95%CIs. Diamond at the bottom shows the pooled effect estimate. Ten randomized controlled trials comparing five potassium-competitive acid blockers [vonoprazan, tegoprazan, fexuprazan (FXP), keverprazan, zastaprazan] with proton pump inhibitors (PPIs) for erosive esophagitis healing are shown. High-acid dissociation constant (pKa) potassium-competitive acid blockers demonstrated significant superiority, while tegoprazan (pKa of 5.1) and FXP (pKa of 9.04, the “FXP paradox”) showed no significant advantage over PPIs. Overall healing rates were high across all treatments (91%-99%), with ceiling effects limiting differentiation in mild-to-moderate esophagitis. Notably, FXP (pKa of 9.04) showed no significant advantage over PPIs (odds ratio = 0.95) despite high-pKa, illustrating the FXP paradox. P-CAB: Potassium-competitive acid blocker; PPI: Proton pump inhibitor.

Figure 5 The acid dissociation constant-efficacy relationship among potassium-competitive acid blockers. A: Scatter plot of acid dissociation constant (pKa) vs pooled odds ratio (OR) for Helicobacter pylori eradication [n = 4 potassium-competitive acid blockers (P-CABs), Spearman ρ = 0.80, P = 0.20]; B: Scatter plot of pKa vs pooled OR for erosive esophagitis healing (n = 5 P-CABs, Spearman ρ = 0.80, P = 0.104). In both panels, fexuprazan (FXP) (pKa of 9.04) is highlighted as an outlier with proton pump inhibitor-equivalent efficacy despite high-pKa; C: Individual P-CAB efficacy for erosive esophagitis ordered by pKa, illustrating the FXP paradox: Despite pKa of 9.04, FXP shows OR < 1.0. Dashed line represents ordinary least squares linear regression fit. Error bars represent 95%CIs of pooled ORs. A positive pKa-efficacy trend is observed, but FXP (pKa of 9.04) represents a notable outlier with proton pump inhibitor-equivalent outcomes despite high-pKa. OR: Odds ratio; P-CAB: Potassium-competitive acid blocker; pKa: Acid dissociation constant; PPI: Proton pump inhibitor; TPZ: Tegoprazan; FXP: Fexuprazan; VPZ: Vonoprazan; KVP: Keverprazan; ZPZ: Zastaprazan.

Figure 6 Monte Carlo simulation for statistical power estimation. Power curves showing the relationship between sample size [number of potassium-competitive acid blockers (P-CABs)] and statistical power to detect acid dissociation constant-efficacy correlations at different effect sizes (ρ = 0.60, ρ = 0.70, ρ = 0.80, ρ = 0.90). With only 5 marketed P-CABs, power to detect ρ = 0.80 is limited to 30% (horizontal dashed line indicates 80% power threshold); with n = 4 (Helicobacter pylori), power is 30%. These simulations illustrate that the inherent limitation of 5 marketed P-CABs constrains statistical power for correlation testing, emphasizing the importance of mechanistic plausibility and supporting evidence from subgroup analyses. P-CABs: Potassium-competitive acid blockers.

Figure 7 Funnel plots for assessment of publication bias. A: Funnel plot for Helicobacter pylori eradication studies (30 randomized controlled trials, 32 comparisons). Each point represents one study, plotted by effect size (log odds ratio, X-axis) against precision (standard error, Y-axis). Visual inspection shows symmetric distribution around the pooled effect estimate (vertical line). Egger’s regression test was not significant (P = 0.97), indicating no evidence of publication bias; B: Funnel plot for erosive esophagitis healing studies (10 randomized controlled trials). Distribution appears symmetric upon visual inspection. Egger’s test was not significant (P = 0.14); however, interpretation is limited by the small number of studies (n = 10), and funnel plot asymmetry was not visually apparent. Dashed diagonal lines represent pseudo-95% confidence limits. H. pylori: Helicobacter pylori; OR: Odds ratio.