Abstract
Background/Objectives: Congestion drives most heart-failure (HF) hospitalizations, and several strategies now titrate therapy to a directly measured congestion signal rather than to symptoms; whether they differ by the signal measured has not been examined comparatively. Methods: We performed a systematic review and contrast-based network meta-analysis of randomized trials comparing imaging- or device-based congestion-triggered management with usual care. Thirteen chronic-phase trials formed a star network (no closed loops; rankings exploratory) of four strategies, pressure-guided (N1), lung ultrasound (N2), transthoracic lung impedance (N3), and cardiac-implantable-electronic-device (CIED) intrathoracic impedance (N5), with HF hospitalization as the primary outcome and six prespecified sensitivity analyses; four acute trials formed a gap map. Results: Only pressure-guided monitoring reduced HF hospitalization (0.72, 95% confidence interval 0.57–0.90), stable across sensitivities and matching external data. Lung ultrasound did not reduce hospitalization (0.72, 0.48–1.08) or mortality but reduced urgent visits (0.26, 0.13–0.50). The apparent CIED-impedance harm (1.12, 0.77–1.63) reflected an audible-alert trial and dissolved when early-terminated trials were excluded (0.87). Lung impedance gave the largest estimate (0.51, 0.31–0.83) but rested on one trial. No strategy reduced mortality. Conclusions: The congestion signal acted upon, and who receives the alert, track with the effect, although signal and workflow cannot be separated within trials; clinicians should match modality to the intended outcome, as these tools defer hospitalization, not death.