Recombinant tissue plasminogen activator as a modifier of treatment effect in prosthetic valve thrombosis: systematic review and meta-regression

Background

Obstructive prosthetic valve thrombosis (PVT) is a rare but potentially fatal complication. Optimal first-line therapy remains debated. Historically, surgery has been favoured, but advances in low-dose, slow-infusion recombinant tissue plasminogen activator (rt-PA) protocols have improved outcomes with fibrinolysis.

Purpose

To compare surgical and fibrinolytic approaches for PVT using meta-analysis and meta-regression, focusing on how rt-PA-based protocols may modify treatment effects.

Methods

A systematic search of five databases identified studies comparing fibrinolysis and surgery for PVT. Random-effects meta-analyses estimated pooled risk ratios (RRs) with 95% confidence intervals, and mixed-effects meta-regression assessed the influence of rt-PA use on in-hospital mortality, complete valve restoration (without death), and recurrence.

Results

Thirteen studies (12 observational, 1 RCT; 586 fibrinolysis, 714 surgery) from 1978-2023 were included. Streptokinase (SK) was the most common fibrinolytic agent, while urokinase (UK) and rt-PA were less frequent, with rt-PA use increasing in recent studies.

Across all studies, fibrinolysis and surgery had similar in-hospital mortality (RR 0.62 [0.32-1.21] I²=57%), though fibrinolysis showed lower complete valve restoration (RR 2.02 [1.25-3.27] I²=70%). Restricting analyses to rt-PA-based regimens, fibrinolysis significantly reduced in-hospital mortality (RR 0.12, [0.04-0.40] I²=0%) without affecting valve restoration (RR 1.62 [0.64-4.13] I²=82%).

Meta-regression showed that for cohorts with 0% rt-PA usage, the baseline RR for in-hospital mortality was 1.24 [0.68-2.26]. Each 10% increase in rt-PA usage reduced RR by 19% (0.81 [0.70-0.93], P = 0.003), with low-moderate residual heterogeneity (I² = 27.5%) and 67% of between-study variance explained. For complete valve restoration, the baseline RR at 0% rt-PA usage was 2.74 [1.41-5.34], favouring surgery. Each 10% rise in rt-PA usage corresponded to a nonsignificant 7% RR reduction (0.93 [0.84-1.04], P = 0.20), suggesting a trend toward improved valve restoration with fibrinolysis.

Conclusion

Despite historical preference for surgery, modern rt-PA-based fibrinolysis may be comparable and possibly superior for obstructive PVT, providing similar rates of valve restoration while significantly reducing in-hospital mortality. Meta-regression identifies rt-PA as a strong moderator of treatment effect, with greater usage linked to lower mortality.For image description, please refer to the figure legend and surrounding text.