Author + information
- Received September 27, 1999
- Revision received December 16, 1999
- Accepted February 21, 2000
- Published online June 1, 2000.
- Mehmet Özkan, MD∗,
- Cihangir Kaymaz, MD∗,
- Cevat Kirma, MD∗,
- Kenan Sönmez, MD∗,
- Nihal Özdemir, MD∗,
- Mehmet Balkanay, MD∗,
- Cevat Yakut, MD∗ and
- Ubeydullah Deligönül, MD, FACC†,* ()
- ↵*Reprint requests and correspondence: Dr. Ubeydullah Deligönül, Cardiology Division, UTHCT, 11937 US Highway 271, Tyler, TX 75708-3154
We analyzed the results of intravenous thrombolytic treatment under transesophageal echocardiographic (TEE) guidance in prosthetic valve thrombosis.
Thrombotic occlusion of prosthetic valves continues to be an uncommon but serious complication. Intravenous thrombolytic treatment has been proposed as an alternative to surgical intervention.
In a four-year period, 32 symptomatic patients with prosthetic valve related thrombosis underwent 54 thrombolytic treatment sessions for the treatment of 36 distinct episodes. All patients had low international normalized ratio values at the presentation. Transesophageal echocardiography was performed at baseline and repeated after each thrombolytic treatment session (total 98 TEE examinations). Streptokinase was used as the initial agent with a repeat dose given within 24 h when necessary. Recurrent thrombosis was treated either with tissue plasminogen activator or urokinase.
The initial success after first dose was only 53% (17/32) but increased up to 88% (28/32) after repeated thrombolytic sessions upon documentation of suboptimal results on TEE examination (p < 0.01). In addition, four asymptomatic patients with large thrombi were also successfully treated with single infusion. The TEE characteristics of thrombus correlated with clinical presentation and response to lytics. Success was achieved with single lytic infusion in 40% of the obstructive thrombi as compared with 75% of the nonobstructive ones (p < 0.05). The success rates of lytic treatment were similar for mitral versus aortic valves, and for tilting disk versus bileaflet valves. Rapid (3 h) and slow (15 to 24 h) infusion of streptokinase resulted in similar success rates. However, major complications (three patients) occurred only in the rapid infusion group.
In patients with prosthetic valve thrombosis, intravenous slow infusion thrombolysis given in discrete, successive sessions guided by serial TEE and transthoracic echocardiography can be achieved with a low risk of complications and a high rate of success.
Thrombotic prosthetic valve occlusion is an uncommon but serious complication that has been reported to occur in 0.5% to 8% of the left-sided mechanical prosthetic valves and in up to 20% of tricuspid prostheses (1–3). Surgical thrombectomy and valve replacement have been used routinely for this condition, although the operative mortality was quite high in certain subsets (4,5). Intravenous lytic treatment for thrombosed prosthetic valves has been used as an alternative to surgical thrombectomy and valve replacement (6–31). In recent years, transesophageal echocardiography (TEE) has greatly improved the detection of intracardiac thrombus and valve dysfunction (32–34). Likewise, several authors (35–38) have recommended the use of TEE in the management of thrombotic prosthetic valve occlusions. However, the indications and delivery methods for thrombolytic treatment in prosthetic valve thrombosis have not been well defined. Recently, new treatment algoritms have been proposed in two publications (29,30). The new American College of Cardiology/American Heart Association (ACC/AHA) guidelines on valvular disease (39) essentially adopted the approach proposed in these articles, with more emphasis on the potential side effects of thrombolysis. The knowledge base for the thrombolytic treatment of prosthetic valve thrombosis is mostly derived from heterogeneous patient series and anecdotal evidence. Therefore, the questions remain regarding the selection of patients according to the thrombus characteristics, the duration and dose of the thrombolytic treatment, the proper monitoring method and the role of intensified anticoagulation as compared with early thrombolysis.
In the present study, we analyzed the results in a consecutive series of patients with prosthetic heart valves in whom the thrombolytic treatment was administered in discrete, successive treatment sessions guided by serial TEE. The success and the complication rates were analyzed according to the echocardiographic and clinical characteristics of the patients.
The study group included 32 consecutive symptomatic patients with mechanical heart valves (13 male, 19 female; average age, 36 ± 12 years) who underwent 54 intravenous thrombolytic treatment sessions in the same, single institution between May 1993 and June 1997 for the treatment of 36 distinct valvular thrombotic events (one involving two valves in the same patient). In 6 of these 32 patients, an additional left atrial mural thrombus was also present. These events involved either hemodynamically significant thrombotic valvular obstruction or large valvular thrombus formation with embolic events despite normal valve function (Table 1). In addition, 4 asymptomatic patients (2 male, 2 female) with mechanical heart valves and large (>10 mm in diameter) thrombi were analyzed separately, thus bringing the total number of patients to 36.
The TEE was performed using 5-MHz multiplane transducer connected to a system (Vingmed CFM 800) after oropharyngeal anesthesia (10% lidocaine) and conscious sedation (intravenous midazolam, 1 to 3 mg). Transthoracic echocardiography (TTE) to supplement gradient and valve area calculations was performed using the same system with 3.25-MHz transducers. All echocardiographic examinations were performed at baseline and repeated after each thrombolytic treatment session (total 98 TEE examinations). Each recording was initially evaluated by two echocardiographers independently and the consensus by both reviewers was achieved in a subsequent session. The protocol was approved by the Institutional Review Board. Written informed consent was obtained from each patient before entry into the study.
The echocardiographic data used in this study were as follows: prosthetic valve thrombus was recognized as soft and homogeneous, mobile or fixed echo densities located at the valve occluder and/or valve struts. The largest diameter of the thrombus as well as the length of the mobile portion, if present, were measured. A diagnosis of pannus formation was made when fixed, bright echodense structures, sometimes containing focal calcific deposits, were present primarily along the valve ring with extension into the valve orifice. Patients with valve dysfunction due to pannus formation were not included in this series (four patients in the same time period, all confirmed at surgery).
Although different makes of mechanical valves will have different “normal gradient” ranges, we arbitrarily selected the cutoff points for valvular obstruction. These limits were largely based on the previously published values (40). Since we did not have a baseline recording for each valve, more specific cutoff points could not be determined. The limitation of the occluder movement was evaluated subjectively. A significant narrowing of the prosthesis was diagnosed when the Doppler mitral valve area was ≤1.5 cm2 and the mitral valve mean gradient was ≥10 mm Hg, or when the aortic mean gradient was ≥40 mm Hg.
Patients with obstructive thrombus, and those with nonobstructive thrombus, who either had a history of embolization or had a large thrombus mass (≥10 mm base diameter and/or ≥5 mm mobile segment length) were accepted as candidates for thrombolysis. Patients with large thrombus were not excluded. Thrombolytic treatment was contraindicated in patients with bleeding tendency and in those with expanding or hemorrhagic cerebral infarcts.
Streptokinase (SK) was the initial agent infused in all patients. No loading dose of SK was used. In the early experience (15 patients), a total of 1.5 million units of SK was administered in 3 h. In subsequent patients, a slow infusion of 60,000 to 100,000 U/h for a total of 15 to 24 h was given (Table 2). When the persistence of thrombus was documented by repeated TEE, another infusion of 1.5 million units of SK was given. Urokinase (UK) or tissue plasminogen activator (tPA) (depending on the availability) was chosen for recurrent thrombus or in the case of failure of two subsequent SK sessions. The tPA dose was 100 mg (10-mg bolus followed by the infusion of the remaining amount in 5 h). The UK dose for each session was 1.5 million units given as a 15-h continuous infusion (100,000 U/h).
Definition of thrombolytic success
The response to thrombolytic treatment was defined as a complete success when significant narrowing of the valve (based on hemodynamic measurements mentioned above) was no longer present and a ≥75% reduction in largest diameter of the thrombus mass was achieved. For nonobstructive thrombi, a reduction by ≥75% in thrombus diameter or complete lysis of the mobile portion of the thrombus was required as criteria for complete success. The positive responses that were less than “completely successful” were accepted as partial success.
Heparin and warfarin treatment was started in patients with complete and partial success and heparin therapy was continued until an international normalized ratio (INR) of ≥2.5 was achieved.
Chi-square test and Student’s t test were utilized for the comparison of discrete and continuous variables, respectively. A p value of <0.05 was accepted as significant.
The types of the prosthetic valves and clinical status at presentation were summarized in Table 1. Of the 32 symptomatic patients, 23 had mitral, 7 had aortic and 1 had tricuspid valve prostheses. In one patient thrombus complicated both aortic and mitral prostheses. The etiology of valve disease was rheumatic in all patients. The types of prosthetic valves were 13 tilting disks and 11 bileaflet valves in mitral position, and 5 tilting disks and 3 bileaflet valves in aortic position. The only tricuspid prosthesis in this series was a bileaflet design. The average interval between valve replacement and presentation was 133 ± 132 weeks (1 to 624 weeks; median, 103 weeks). In 4 of 32 patients (12.5%), recurrent thrombus was documented 20 ± 6 weeks (median, 20 weeks) after initially successful thrombolysis. One of these four patients had a third recurrence 22 weeks after the last one. Of note, all patients had subtherapeutic INR values at the presentation (1.3 ± 0.3). Recurrent thrombotic events were also associated with low INR values. Atrial fibrillation was present in 19 of 36 patients (53%).
The symptoms at presentation (Table 1) of 36 distinct clinical episodes in 32 patients were mainly dyspnea (58%) and embolic events or transient ischemic attacks (36%). Interestingly, four patients reported muffling or cessation of their prosthetic valve sounds. The New York Heart Association (NYHA) functional class was III or greater in 50% of the episodes.
The number and the type of treatment sessions are also shown in Table 1. In the mitral position, 41 thrombolytic sessions (including 4 recurrent thrombus) were performed in 27 episodes in 23 patients. In aortic position 9 thrombolytic sessions were given for 7 episodes in 7 valves. The patient with both aortic and mitral valve thrombi was given SK in two successive sessions. The patient with tricuspid prosthesis previously had SK infusion at another institution and was presented with recurrence (14). Doppler hemodynamic measurements before and after the final session of thrombolytic treatment were listed in Table 2.
The outcome of thrombolytic treatment
According to the study definitions, four episodes resulted in partial success (Table 3). These were included in the complete success group for their small numbers. The initial rate of success after the first dose of SK in primary valvular thrombosis was only 53% (17/32), but increased up to 87.5% (28/32) after repeated thrombolytic sessions upon documentation of suboptimal results on TEE examination (p < 0.01). One additional thrombolytic session was administered in 17 episodes. A total of three thrombolytic sessions were necessary in only one patient, who had an unsuccessful result. In four patients with five episodes of recurrent thrombosis, a single dose was successful in two and two successive treatments were necessary in the remaining three. Of the three patients with no response to lytics, two were maintained on regimens of higher-dose oral anticoagulants and the third one underwent surgery.
The initial and overall success rates of thrombolytic treatment in mitral and aortic valves were 50% and 63%, and 87.5% and 87.5%, respectively (p = NS). The overall success rates in tilting disk and bileaflet valves were 86% and 94%, respectively (p = NS). The success rates for thrombolysis were 89%, 92% and 80% in patients with NYHA class ≤II, III and IV symptoms, respectively (p = NS). There was no difference in success rate for rapid as compared with slow infusion of SK (Table 4).
Echocardiographic thrombus characteristics and thrombolytic response
Baseline TEE results for each valve in the symptomatic initial (n = 32) and recurrent (n = 5) thrombotic episodes were analyzed together (total 37 valve-episodes) (Table 5). Interobserver agreement was achieved in 96 of 98 TEE recordings (98%) regarding the diagnosis of thrombus and definition of its characteristics. Thrombus characteristics were correlated with clinical presentation. A fixed thrombus mass measuring at least 10 mm at its base was documented in 21 episodes (57%) of which all were associated with hemodynamically significant valvular obstruction. In contrast, a mobile thrombus with significant motion of the thrombus mass (average length, 12.1 ± 5.5 mm; 5 to 22 mm), regardless of the size, was found in 16 episodes (43%), and it was associated with valvular obstruction in only 4 (25%, p < 0.001). However, mobile thrombi were found in 10 of 13 discrete episodes that were presented with embolic events (77%), while only 6 of 24 (25%) episodes without embolism were associated with mobile thrombi (p < 0.05). Nonobstructive thrombi were smaller (8.6 ± 5.3 mm vs. 17.6 ± 7.6 mm, p < 0.001) and they were less likely to be associated with class III or greater symptoms at presentation as compared with the obstructive thrombi (8% vs. 71%, p < 0.001).
After the initial thrombolytic dose, complete success was achieved in 10 of the 25 (40%) obstructive thrombi. In nine more instances, thrombus was lysed enough to become nonobstructive and in the remaining six, it continued to be obstructive despite some (n = 4) or no (n = 2) change in appearance. Complete success was confirmed in 12 of these 15 episodes after repeated administration of thrombolytics, bringing the total success rate to 88% (22/25). Among the 12 nonobstructive thrombi, however, complete success was achieved in 9 (75%) at the initial thrombolysis attempt (p < 0.05, as compared with obstructive thrombi). The remaining three patients with nonobstructive thrombi were given a second dose of thrombolytics with success in two, thus bringing the overall success in this subset to 91.6%. The overall success rates in fixed (90.5%) and mobile (87.5%) thrombi were not statistically different.
Thrombolytic treatment in asymptomatic patients
There were two male and two female patients with mitral valve prostheses (three bileaflet, one tilting disk) who underwent successful thrombolysis for large, nonobstructive left atrial thrombi detected on routine examination (Table 1). The thrombi were mobile in two and fixed type in the remaining two. A single SK infusion session was successful in all four patients.
Left atrial mural thrombus associated with thrombosis of the mitral valve prosthesis
The left atrial mural thrombus persisted in six patients despite successful lysis of the thrombus associated with the prothesis itself. No systemic embolization was noted in this group.
Complications of thrombolytic treatment
The symptomatic and asymptomatic groups of patients combined for the analysis of complications. The complications of thrombolytic treatment were seen in 15% of the 40 total thrombotic episodes in 36 patients (Table 1). Major complications included death in one patient (2.8%), coronary embolization in one patient (2.8%) and cerebral embolization in one patient (2.8%). Both patients with embolization had large, fixed, obstructive thrombi on TEE. Death was due to cerebral bleeding in a patient with 10-day-old brain embolism with no evidence of hemorrhagic infarct on baseline computed tomographic (CT) scan. Minor bleeding was noted in three (8.3%) of the episodes. Of interest, all major complications occurred in the group that received the rapid infusion of SK as initial treatment. No complications occurred in four asymptomatic patients. No complications occurred as a result of TEE examinations.
Our study showed that intravenous thrombolysis can be achieved with a high success rate in all echocardiographic subsets of prosthetic valve thrombosis by administering thrombolytics in successive episodes. In slightly more than half of the events, a single infusion of thrombolytic was successful, and one additional infusion ensured a successful result in about three-fourths of the remaining. Additional infusions were required more often in patients with obstructive as compared with nonobstructive thrombus. The relief of hemodynamic obstruction by thrombolysis can be monitored simply by transthoracic Doppler echocardiography. Fluoroscopy may also be useful by detecting the return of valve occluder motion to normal range (22). In addition to its fundamental value in detecting nonobstructive thrombus, TEE may still provide additional valuable information in other subsets of prosthetic valve thrombosis that may be important in the treatment of individual patients. Although the changes in thrombus mass generally mirrored the hemodynamic changes in patients with obstructive thrombi, in 4 of these 25 patients, initial thrombolysis caused reduction in the thrombus mass with no change in hemodynamic obstruction. These four patients received additional thrombolytic infusions with relief of obstruction. In three patients with no initial change in thrombus configuration by TEE after the first thrombolytic dose, subsequent doses also failed. Potentially, TEE may be useful in identifying these “resistant” cases. Nevertheless these are observations in small groups and the potential impact of serial TEE in this context deserves to be investigated in larger series.
Obstructive versus nonobstructive thrombus
Although the hemodynamic benefit of lysing an obstructive thrombus is obvious, the issue with nonobstructive prosthetic valve thrombus is somewhat controversial. Recently published guidelines by Lengyel et al. (30) proposed intensive anticoagulation, rather than thrombolysis, in patients with nonobstructive thrombi. Valvular thrombi without hemodynamic obstruction can be detected with higher accuracy by TEE (35) and may be seen in 15% to 18% of the patients in the early postoperative period following valve replacement (41). Despite the absence of hemodynamic embarrassment, nonobstructive prosthetic thrombi may not be totally benign. We and others showed that nonobstructive mobile prosthetic valve thrombi carried an increased risk of systemic embolization (35,41). In the current study, the thrombolytic treatment was not associated with increased embolic complications in this subset despite the more frequent initial presentation with embolic events. Similarly, a more recent study, again by Lengyel et al. (42), reported that the anticoagulant treatment failed in half of the patients with nonobstructive thrombi, while there was no failure or complication in the group treated with thrombolysis. Together these results bring strong support to extending the indications for thrombolytic treatment to patients with nonobstructive thrombi. Transesophageal echocardiography may show persistent residual thrombus (with embolic potential) following initial thrombolysis, even if the relief of hemodynamic obstruction is detected by other means. In our series, 9 of 25 (32%) obstructive thrombi were transformed into nonobstructive but large thrombi (4 of them gaining mobility) after the first SK dose and were treated successfully with additional infusion of the thrombolytic in all but one.
Thrombolysis results according to the valve type and location
Roudaut et al. (20) reported higher thrombolytic success rates for aortic than mitral prosthetic valves. Some authors suggested that the bileaflet valve thrombotic occlusions were more responsive to thrombolysis (20,23). An explanation for this was the potential for the flow obstruction by a smaller thrombus occupying the critical pivot point in a bileaflet valve as compared with a single disk valve. Our series, in contrast, did indicate similar success rates in both valve locations and valve types. This was true even when the success rates for each valve location and type were further stratified according to fixed or mobile, and obstructive or nonobstructive thrombus characteristics.
Thrombolysis results according to the clinical status
The thrombolytic success was achieved in equal proportion of patients who were in class II or less as compared with those who were in class III or greater. Hurrell et al. (29) recommended direct surgical treatment rather than thrombolytics for patients with class IV symptoms, because of a high medical mortality reported in some series. The ACC/AHA valve disease guidelines (39) also recommended surgical treatment in class IV patients except those who are nonsurgical candidates. Unfortunately, the surgical mortality is also highest in this group of patients (4,5). Recently published thrombolytic therapy guidelines (30) proposed the thrombolytic treatment as the treatment of choice for the “critically ill” patients in functional class III or IV, although the definition of “critically ill” may be quite variable. We believe that an individual approach based on the expected surgical mortality rate is necessary in patients with class IV symptoms. We suggest that, in the absence of cardiogenic shock, these patients may be treated in the intensive care unit by TEE-guided thrombolysis.
Systemic embolization and other complications
Due to the lack of a generally accepted standard regimen, various infusion protocols and thrombolytic doses were used in the previous studies (6–31), making a direct comparison of the results difficult. The rate of major events and minor bleeding in our study was within the previously reported range. The embolization rate of 5.7% (two patients) in our series compares favorably with previously reported single institution series (Table 6). Both embolic complications occurred in patients with fixed, obstructive thrombi who were given rapid thrombolytic infusions. In the recent thrombolytic treatment guidelines article (31), the presence of a large thrombus on the prosthetic valve or in the left atrium has been included in the relative contraindications based on one case report of embolism following a rapid (3-h) infusion of tPA (26). In a recent study by Manteiga et al. (31) 3 instances of embolization (33%) occurred in 10 patients with large left atrial thrombi who were given rapid infusion of SK (1.5 million units in 90 min). It appears that rapid thrombolytic infusion, especially in the presence of large thrombus, may increase the risk of embolization. Understanding the predictors of systemic embolism is important because this complication seems to be the Achille’s heel of the thrombolytic treatment of prosthetic valve thrombosis. There are little data available on this subject. Reddy et al. (24) published the series with lowest embolization rate and suggested that advanced age and atrial fibrillation may be the risk factors for embolic complications. Roudaut et al. (20), however, found that class IV clinical status was the only predictor of embolic complications (and death). It is not known whether this was secondary to larger amounts of thrombi in these patients.
The only mortality in our series was the result of cerebral bleeding in a patient with a 10-day-old cerebral embolism. Although the baseline CT scan had not shown bleeding, a repeated CT before second thrombolytic application might have been important in this case. As a precaution, probably no patient with cerebral embolism within 10 days should be given thrombolytics and each patient should be carefully assessed individually for the risk of hemorrhagic transformation of cerebral infarct.
In this series, intravenous thrombolysis for prosthetic valve thrombosis was achieved with a low risk of complications and a high rate of success in both obstructive and nonobstructive prosthetic valve thrombosis. Thrombolytic treatment was equally effective in both mitral and aortic valve thrombi, and in both bileaflet and single disk valve designs. We believe that the slow administration of intravenous thrombolytics in discrete, successive sessions guided by serial TEE and TTE information is a safe and effective method that may expand the indications for nonsurgical treatment of prosthetic valve thrombosis.
- American College of Cardiology/American Heart Association
- computed tomography
- international normalized ratio
- New York Heart Association
- transesophageal echocardiography
- tissue plasminogen activator
- transthoracic echocardiography
- Received September 27, 1999.
- Revision received December 16, 1999.
- Accepted February 21, 2000.
- American College of Cardiology
- Husebye D.C,
- Pluth N,
- Piehler J.M,
- et al.
- Deviri E,
- Sareli P,
- Wisenbaugh T,
- Cronje S.L
- Luluaga I.T,
- Carrera D,
- D’oliveria J,
- et al.
- Baille Y,
- Choffel J,
- Sicard M.P,
- et al.
- Witchitz S,
- Veyrat C,
- Moisson P,
- et al.
- Ledain L.D,
- Ohayon J.P,
- Colle J.P,
- et al.
- Kurzrok S,
- Singh A.K,
- Most A,
- William D.O
- Enar R,
- Ersanli M,
- Kucukoglu S,
- et al.
- Zoghbi W.A,
- Desir R.M,
- Rosen L,
- et al.
- Mehan V.K,
- Dalvi B.V,
- Kale P.A
- Roudaut R,
- Labbe T,
- Lorient-Roudaut M.F,
- et al.
- Silber H,
- Khan S.S,
- Matloff J.M,
- Chaux A,
- De Robertis M,
- Gray R
- Lengyel M,
- Fuster V,
- Keltai M,
- et al.
- Nanda N.C,
- Cooper J.W,
- Mahan E.F,
- Fan P
- Dzavik V,
- Cohen G,
- Chan K.L
- Gueret P,
- Vignon P,
- Fournier P,
- et al.
- Currie P,
- Sutherland G.R,
- Starkey I.R
- Bonow R.O,
- Carabello B,
- de Leon A.C,
- et al.
- ↵Weyman AE. Principles and Practice of Echocardiography, 2nd ed. Lea and Febiger, 1994; 1210.
- ↵Özkan M, Kaymaz C, Kirma C, Pektas O. Diagnostic approach of mitral and aortic prosthetic valves thrombosis with transesophageal echocardiography in relation to embolic events (abstr). Eur Heart J 1997;18 suppl:P1810;325.
- Lengyel M,
- Temesvari A,
- Nagy A,
- et al.