Discussion
The lower velocity of flow in the right heart and the commonly larger size of TV prosthesis, makes thrombus formation easier. Moreover, RV dysfunction, atrial fibrillation, right atrial enlargement, pulmonary arterial hypertension, hypercoagulable state, and multiple prostheses are other common predisposing factors; therefore, biological valves were assumed as the gold standard and mechanical valves as the challenging device in TV position [8,10-12]. However, mechanical valve has been used more frequently due to its presumed longer durability in the case of adequate anticoagulation maintenance, and higher risk for redo surgery in these patients, as they usually have a history of multiple sternotomies owing to previous TV repair, frequent other valve replacements, or congenital heart surgeries [13,14]. The present report is the largest and the most comprehensive single-center study evaluating the isolated mechanical TV thrombosis with respect to treatment modalities and response to therapies during follow up period.
Our patients have encountered this serious and morbid condition after an average elapsed time from surgery of 49.9 (range 0.4-276) months, while 29.8% of episodes occurred in ≤12 and 46.2% in ≤24 months following surgical TV replacement. This is almost in agreement with previous studies evaluating the rate and prognosis of TV thrombosis. Roudaut et al. [11] described a median time from valve replacement to thrombosis of 4.3 years (14 days to 18 years) and an incidence rate of thrombosis in 24% of patients in the first postoperative year. In a meta-analysis by Liu et al. [10] valve thrombosis occurred at a mean follow-up of five to seven years and structural bioprosthetic degeneration occurred at an average of seven to nine years after TV replacement. These findings emphasize the importance of adequate anticoagulant treatment with a special consideration of postoperative period as well as during follow-up period.
The sub-therapeutic INR values less than 3 and 3.5 was seen in 70% and 82% of episodes at the time of admission, respectively. As it could not be the only clue to anticoagulation status, due to a probable more chronic process of a usually multilayered clot, we also assessed the average amount of previous 3-6 months. Contrary to left-sided valves, guidelines do not specify recommended levels of INR for mechanical prosthesis in TV position. It is advocated that target INR should be adjusted to patient risk factors and the thrombogenicity of the prosthesis (i.e., mitral or tricuspid position, atrial fibrillation, and a history of previous thromboembolism). In addition, a single INR target is mandatory for efficient antithrombotic therapy to identify that the acceptable range is 0.5 units of INR levels on each side, avoiding the upper or lower edge of the range values. We could not determine such a target due to our study type, but according to guidelines implications and considering the valve type in our study, we recommend the INR target of at least 3 for lower risk patients and 4 for higher risk ones with a history of previous thrombosis [3,4] Shapira et al. [15] recommended the target level of INR as 3.5-4 for patients with TV prostheses, and Zhang et al. [16] considered the target INR levels of 3-3.5. On the other hand, we should also take into account that patients may have lesser compliance with higher INR goals [17]. In addition, Aspirin 75 to 100 mg daily in combination with vitamin K antagonist anticoagulation in all patients, and the continuation of warfarin with a therapeutic INR levels in patients undergoing minor procedures is recommended to prevent the rapid decrease in the INR and the risk of thromboembolism [3,4].
Complete hemodynamic and clinical success rate of 68.2% following thrombolytic therapy was observed in this study. It is fairly within the ranges of previous similar studies [7,9,14,18,19]. It should be considered that different success rates may be possibly caused by different patient selection, diagnostic modalities, fibrinolytic regimens [20], and the definition of outcomes. Additionally, most of our cases (97%) were diagnosed as obstructive TV thrombosis, and we evaluated successful response in absence of important complications (such as retroperitoneal hematoma and stroke). There was no statistically significant difference regarding the success rate among thrombolytic treatment subgroups. This might be due to insufficient number of cases in certain subcategories [20].
Our study highlighted that thrombolytic response depends on the chronicity of thrombosis and the presence of pannus formation. We should bear in mind that valve thrombosis could be an acute, subacute or chronic event and thrombi are typically formed of different clot layers, with variable degrees of organization. This indicates that trying thrombolytic therapy as the first step treatment is sensible in most patients, even though probably with an incomplete resolution of clot [11,12]. In addition, regarding the treatment strategy in recurrent cases, it seems that the prior management of patients had a substantial impact on treatment choice by physicians. In those with a history of prior thrombolysis and anticoagulant intensification there were more trends towards surgery, while in patients with previous TV surgery, more patients were treated with fibrinolytic. In a former study in cases received fibrinolytic, it was revealed that the level of anti-tissue plasminogen activator antibodies (ATA) may increase during thrombolysis with rt-PA infusion and may interfere with the success rate, necessitating a higher dose for complete success and play a role in resistance to fibrinolysis [21].
Retroperitoneal hematoma may happen unexpectedly in patients under anticoagulant treatment. It may be life-threatening and leads to abdominal compartment syndrome, requiring prompt surgical approach [22]. The lowest non-cerebral major bleeding was reported to be 1.7% by Ozkan et al. [23] with ultraslow thrombolytic therapy. Moreover, we found an embolic cerebrovascular event during thrombolytic therapy, in spite of being primarily unpredictable in right sided PVT fibrinolysis. Cerebral embolism following thrombolytic treatment have been reported to have an incidence of 3-10% and is even more frequent in the presence of atrial fibrillation [20]. This heighten the role of TEE in risk stratifying of patients for thrombolysis. Given the high rate of left-sided valve replacements, congenital heart disease and AF rhythm among TVR patients, they should be evaluated for the presence of thrombus formation on left-sided prosthetic valves, left atrium or its appendage, and probable PFO as the source of emboli, as well.
Moreover, despite the low rate of mortality and major non-fatal complications in the course of in-hospital management, there were considerable recurrent episodes throughout the entire follow-up period. Thirty-eight percent of our patients experienced at least one episode of recurrent TV thrombosis. Recurrence rates after thrombolytic therapy were reported from 11% to 32% in similar studies [18,21,24] and fibrinolysis was associated with a significantly higher rate of recurrence, though none of previous reports has described in isolated mechanical TV thrombosis [18] According to our findings, it seems that incomplete resolution of thrombosis and the presence of pannus are remarkable triggers for re-thrombosis. We should mention the “Doppler silent” phenomenon, which implies that complete or acceptable normalization of valve occluder movement on fluoroscopy and valve gradients on echocardiography may not be merely the conclusive signs of complete thrombus resolution, and further visualization by TEE seems to be essential [25] So, a considerable number of our patients might have needed the continuation of thrombolysis as TEE was not performed in a significant number of patients.
Finally, novel diagnostic and therapeutic approaches are still evolving. Sixty-four-slice multi-detector computed tomography proved to be helpful in identifying masses amenable to thrombolysis in patients with prosthetic valve dysfunction. A high attenuation suggests pannus overgrowth and is associated with reduced response to fibrinolysis, whereas a lower value is associated with thrombus formation [26]. With regard to treatment, catheter‐directed thrombolysis in selected patients may render the potential advantage of local thrombolysis with a much lower dose than systemic therapy, possibly providing better efficacy while reducing the risk of bleeding complications [8,16,27]. The selection of type of prosthetic valve is another important issue, so that the availability of transcatheter valve-in-valve replacement can change the trade-offs between mechanical and bioprosthetic valves [3]. Tricuspid valve-in-valve procedure was shown to be technically feasible in a number of studies in patients with a wide range of valve size [13,28]. This will be more interesting when we consider the relatively short average time to thrombosis and a high recurrence rate in our study together with the data derived from former studies indicating mean reoperation time of 44.2 months for mechanical prosthesis [29].