Prosthetic valve thrombosis is an uncommon complication, with incidences reported to be below 6% for mechanical valves and under 1% for biological valves, with the mitral position carrying higher risk compared to the aortic position. In cases of postcardiotomy shock, circulatory support with VA-ECMO might be required, some of whom have prosthetic valves. These cases might present an increased risk of prosthetic thrombosis due to the hemodynamic alterations induced by ECMO. Several case series have been published, though there is no specific evidence or recommendations regarding its management.
The goal of the article under review is to describe the incidence, management, and outcomes of patients who developed prosthetic thrombosis while on VA-ECMO support after valve replacement surgery. This study was conducted at the Pitié-Salpêtrière Hospital in Paris, exclusively including patients from this center. Inclusion criteria encompassed adult patients who received VA-ECMO via peripheral access after valve replacement surgery between January 2015 and October 2019. Indications for VA-ECMO support were based on clinical criteria of postcardiotomy cardiogenic shock. Peripheral femoro-femoral cannulation was performed in all cases, and some patients received intra-aortic balloon pump (IABP) support due to echocardiographic findings of left ventricular dysfunction and/or limited opening of the aortic valve/prosthesis. ECMO flow was adjusted according to contractility and cardiac output, with daily echocardiographic assessments. Anticoagulation was maintained with continuous unfractionated heparin infusion, targeting an activated partial thromboplastin time (aPTT) ratio of 1.5–2. Cannulas and oxygenators were inspected daily by perfusionists, and replacements were performed if thrombi were detected around the membrane.
The primary objective was to determine the incidence of prosthetic thrombosis during VA-ECMO support or within days following weaning. Secondary objectives included hospital and 30-day survival, as well as adverse events related to peripheral VA-ECMO in these patients.
From 2015 to 2019, 1936 patients underwent valve replacement surgery, of whom 152 (7.8%) required VA-ECMO. Among them, 69/152 (45%) received combined IABP and VA-ECMO support. Prosthetic valve thrombosis occurred in 9/152 patients. Of these, 7 underwent biological aortic valve replacement, and 2 underwent double valve replacement with mechanical aortic and mitral prostheses. Five cases developed thrombosis within the first 24 hours, while the remaining four occurred between days 4 and 17. The cumulative incidence of prosthetic thrombosis was lower in patients with IABP and VA-ECMO compared to VA-ECMO alone (1.4% vs. 13.7%; p = .021). Embolic events (peripheral and cerebral) were more frequent in patients with prosthetic thrombosis compared to those without (22.2% vs. 4.2%; p = .02). No statistically significant differences were found regarding prosthesis position and embolic events. Survival among patients with prosthetic thrombosis was 22% (2/9), compared to 31% in those without thrombosis. Among the nine cases, five were managed by converting to biventricular bicentrifugal support (Abbott Centrimag®). The two survivors were treated with unfractionated heparin infusion without additional surgery.
The authors identified the main limitations of their study as the inability to perform multivariable analysis due to the low number of cases (n = 9) and the single-center design, which limits external validity. They concluded that peripheral femoro-femoral VA-ECMO combined with IABP is associated with a lower risk of prosthetic thrombosis, and this finding appears to be independent of prosthesis type or position.
COMMENTARY:
This article presents the largest case series to date of prosthetic valve thrombosis in patients on femoro-femoral VA-ECMO support. The study identifies two thrombosis patterns: early (within the first 24 hours) and late. The primary hypothesis for early thrombosis involves polytransfusion of blood products and prothrombotic agents immediately after surgery, while late thrombosis could be explained by the inflammatory response induced by prolonged ECMO support and the increased afterload generated by peripheral VA-ECMO.
Another notable finding is the lower cumulative incidence of prosthetic thrombosis in patients managed with combined VA-ECMO and IABP compared to VA-ECMO alone. The retrograde flow generated by femoro-femoral VA-ECMO in a myocardium compromised by postcardiotomy shock is detrimental to recovery, and the authors advocate for the combined VA-ECMO + IABP strategy whenever possible.
As for recommendations, the initial management of chronic-subacute prosthetic valve thrombosis usually requires valve replacement surgery or thrombolysis. Extrapolating this scenario to patients with VA-ECMO in postcardiotomy shock is not considered viable by the authors. Instead, they propose alternatives such as converting to biventricular support to maintain intracavitary flow and enhancing anticoagulation with unfractionated heparin infusion.
A limitation of the study lies in the choice of peripheral arterial cannulation, justified by a 2020 study concluding that high-flow central arterial cannulation in the immediate postoperative period leads to intracardiac stasis and increases thrombosis risk. Among peripheral cannulation options, the femoral artery is preferred over the axillary artery, as the latter has been associated with higher risks of accidental decannulation and bleeding.
Peripheral VA-ECMO offers clear advantages over central VA-ECMO, such as reduced risks of mediastinitis and bleeding with a closed chest, as well as better ventilation capacity, minimizing the risk of atelectasis.
Regarding prosthesis position and its relationship to peripheral embolisms, the authors highlight previous series reporting higher embolism rates in mitral prostheses. For instance, a cohort of patients with peripheral VA-ECMO showed a 21% embolism rate among mitral prosthesis carriers, contrasting with the 7% reported in this cohort. However, these are case series with insufficient numbers to generate strong evidence. Therefore, it seems premature for the authors to assert that variables like prosthesis type and position do not influence embolism genesis.
Although this study features a relatively large cohort, it remains insufficient to establish strong, reproducible recommendations for all cardiac surgery centers. A systematic review of cases from multiple centers is likely the best long-term solution, enabling the development of unified criteria for managing this rare but critical complication, which remains a significant concern for critical care units handling VA-ECMO patients.
REFERENCE:
Danial P, Zamorano C, Carillion A, Barreda E, Laali M, Demondion P, et al. Incidence and outcomes of prosthetic valve thrombosis during peripheral extracorporeal membrane oxygenation. Eur J Cardiothorac Surg. 2024 Nov 4;66(5):ezae321. doi: 10.1093/ejcts/ezae321.
