Surgery or Fibrinolysis for Thrombosed Mechanical Valves

By Michael H. Crawford, MD, Editor

Synopsis: A randomized trial of urgent surgery vs. thrombolytic therapy for symptomatic left-sided mechanical valve thrombosis has shown that surgery completely restores valve function in all patients compared to three-quarters of patients with thrombolytic therapy, but at a cost of higher in-hospital complications and mortality.

Source: Karthikeyan G, Rajashekar P, Devasenapathy N, et al. Urgent surgery vs fibrinolytic therapy for left-sided prosthetic valve thrombosis: A randomized trial. Eur Hear J. 2025; Jun 27. doi: 10.1093/eurheartj/ehaf391. [Online ahead of print].

Previous observational studies of symptomatic patients with thrombosed left-sided mechanical prosthetic valves have shown that urgent surgery completely restores valve function but has a high operative mortality rate. Low-dose slow infusions of tissue plasminogen activator (tPA) appear to have a higher likelihood of restoring valve function with lower complication rates compared to streptokinase-based thrombolytic therapy. Because of the absence of randomized trials, these investigators from a tertiary center in New Delhi, India, conducted a randomized trial of symptomatic patients with mechanical prosthetic valve thrombosis (PVT) comparing urgent surgery to fibrinolysis with tPA (SAFE-PVT).

The study was open-label, but outcomes were blindly adjudicated by an independent committee. Adults with left-sided mechanical prosthetic heart valves with symptoms and objective evidence of prosthetic valve dysfunction for ≤ 2 weeks were recruited. Excluded were patients with contraindications to thrombolytic therapy, intracerebral hemorrhage, active bleeding, recent ischemic stroke, left atrial thrombus, pregnancy, or those deemed too high a surgical risk. Randomization was stratified by New York Heart Association (NYHA) class.

In the study, tPA (25 mg) was infused over six hours and repeated if necessary up to 150 mg. The decision to replace the valve or perform thrombectomy was left to the individual surgeon’s discretion. The primary endpoint was hospital discharge with normal valve function without a stroke, major bleed, or systemic embolism. Secondary endpoints included mortality, recurrent valve thrombosis, stroke, systemic embolism, major bleeding, or continued valve dysfunction over a one-year follow-up. The analysis was by intention-to-treat.

Eighty patients were estimated to be adequate to detect a difference in outcomes with at least 80% power and a type one error set at 0.05. Between 2017 and 2023, 79 patients were enrolled (mean age 36 years, 40% women). Of these patients, 39 were randomized to surgery within a median of 44 hours and 40 to fibrinolytic therapy. The median dose of tPA was 75 mg and six patients received ≥ 150 mg. The primary outcome was achieved in 72% of patients randomized to surgery and 68% of patients receiving tPA (odds ratio [OR], 1.22; 95% confidence interval [CI], 0.47 to 3.19; P = 0.69). All the surgical patients had complete restoration of valve function, as did 73% of the thrombolytic patients. The secondary safety outcome occurred in 11 of the 39 surgical patients and four of the 40 tPA patients (OR, 5.14; 95% CI, 1.28 to 20.59; P = 0.021), which mainly was driven by the seven deaths in the surgical patients as the result of refractory cardiogenic shock vs. one patient in the thrombolytic group as the result of intracranial hemorrhage.

The results did not differ by NYHA class. Two patients in each group had recurrent valve thrombosis during the one-year follow-up; all four died. Stroke occurred in four patients in the tPA arm and one patient in the surgical arm after one year. The authors concluded that in symptomatic patients with left heart mechanical prosthetic valve thrombosis, urgent surgery was no better than fibrinolytic therapy and was associated with a higher rate of complications. A low-dose, slow infusion of tPA was associated with residual valve dysfunction in one-quarter but was much safer.

Commentary

Mechanical heart valves are recommended for younger patients since they have superior durability compared to biological valves and could last the patient’s lifetime. However, they require lifelong oral anticoagulation with vitamin K antagonists to prevent valve thromboembolic complications.

When the international normalized ratio (INR) is maintained in the therapeutic range, thromboembolic complications are uncommon, and valve thrombosis is rare. In health resource-constrained environments where the quality of oral anticoagulation care is suboptimal, thrombotic complications are more common. For example, a retrospective study from India found that symptomatic left-sided PVT occurred in 13% of patients over five years.¹ In areas where access to high-quality heart surgery is limited, thrombolysis, especially with tPA, has become the treatment of choice. On the other hand, guidelines recommend surgery as first-line therapy, with thrombolysis as a reasonable alternative in selected cases based mainly on studies with streptokinase. Therefore, a randomized trial comparing tPA to surgery is of interest.

SAFE-PVT is the first randomized trial of these two approaches, and it showed that surgery completely restored valve function but is associated with a high incidence of in-hospital major complications and death (18%). Thrombolysis restored valve function without embolic complications or major bleeding in two-thirds of patients with PVT with only one death (2.5%). However, the residual valve dysfunction in the thrombolysis group led to four subsequent strokes in the one-year follow-up period compared to one stroke in the surgical group. One might conclude that there is no completely satisfactory therapy for left-sided mechanical PVT. Thus, considerable physician judgment still is required to select the optimal therapy for each patient.

There are limitations to SAFE-PVT. Although the investigators almost hit their predetermined study population of 80 patients, this is a small study, and the results showed wide confidence intervals. Differences in clinical variables not accounted for in the study could influence the results. Transesophageal echocardiography was not done, which could influence clinical decisions. Urgent surgery was done a median of 44 hours post-randomization, and the target of ≤ 48 hours was only achieved in 53%. Also, many patients were turned down by the surgeons, which was one of the exclusion criteria. Thus, the results are confined to a lower surgical risk population and may not be widely applicable. However, the authors suggested that, based on this experience, thrombolysis, not urgent surgery, should be the treatment of choice.

Michael H. Crawford, MD, is Professor Emeritus of Medicine and Consulting Cardiologist, UCSF Health, San Francisco.

Reference

1. Talwar S, Mathur A, Choudhary SK, et al. Aortic valve replacement with mitral valve repair compared with combined aortic and mitral valve replacement. Ann Thorac Surg. 2007;84(4):1219-1225.