Author + information
- Received March 13, 2003
- Revision received August 19, 2003
- Accepted September 23, 2003
- Published online April 7, 2004.
- Guillaume Laplace, MD*,* (, )
- Stéphane Lafitte, MD, PhD*,
- Jean-Noèl Labèque, MD*,
- Jean-Marie Perron, MD*,
- Eugène Baudet, MD*,
- Claude Deville, MD*,
- Xavier Roques, MD* and
- Raymond Roudaut, MD, FESC*
- ↵*Reprint requests and correspondence:
Dr. Guillaume Laplace, Service de Cardiologie, Pr Roudaut, Hôpital Cardiologique du Haut-Lévêque, Avenue Magellan, 33604 Pessac, France.
Objectives The aim of this study was to evaluate the incidence of early thrombosis, its prognostic significance, and the therapeutic implications.
Background Transesophageal echocardiography (TEE) is the method of choice for detecting symptomless early postoperative thrombosis of prosthetic valves. However, the clinical significance is not yet known.
Methods Between June 1994 and December 2000, 680 consecutive patients underwent TEE on day 9 after mechanical mitral valve replacement, to search for early thrombosis. Initially, end points were the in-hospital outcome and treatment. Patients were also evaluated 34 ± 22 months after surgery.
Results Sixty-four early thrombi were detected (9.4%). Two early obstructive were treated by redo-surgery. Sixty-two nonobstructive benefited from medical treatment. The patients were allocated into two groups as a function of the maximum size of thrombus: <5 mm in 29 patients (group A) and ≥5 mm in 35 (group B). During early follow-up, we observed one complicated course in group A and eight in group B. In the long-term survey, three complications were noted in group A and 11 in group B. Incidence of early (p = 0.027) and long-term (p = 0.04) complications were significantly different in the two subsets.
Conclusions This study confirms the incidence of early thrombi after mechanical mitral valve replacement detected by TEE. A small (<5 mm) nonobstructive thrombus seems benign, and our experience argues in favor of medical treatment. Prognosis appears more serious for large thrombi. Medically aggressive therapy and further surgery should be considered in cases of obstructive thrombosis or large and mobile nonobstructive thrombosis.
A number of recent studies with transesophageal echocardiography (TEE) have identified a high incidence of nonobstructive thrombi (NOT) in the early postoperative period after mechanical mitral valve replacement (MMVR) (1–5). Such phenomena are known to increase morbidity and mortality and are often the consequence of inadequate anticoagulant therapy in the highly unstable period after surgery (3,6,7). Transesophageal echocardiography is currently recognized as the method of choice for detecting abnormal echoes, especially those from NOT.
The purpose of our study was to define the incidence of nonobstructive and obstructive thrombosis after implantation of bileaflet mitral prostheses, the optimal management, and the short- and long-term complications of these thrombi.
A total of 680 consecutive patients who benefited from MMVR with a bileaflet prosthesis (Saint Jude Medical, MIRA, ATS, SORIN) from June 1994 to December 2000 in our institution were included in this prospective, monocentric study. All patients underwent systematic, early TEE on day 9 after surgery. The postoperative anticoagulant treatment consisted of intravenous heparin started 6 h after implantation and adjusted from the activated partial thromboplastin time (aPTT). This was followed at the 24th h by subcutaneous heparin three times a day, which was prolonged until the target international normalized ratio (INR = 2.5 to 3.5) was obtained with oral anticoagulant therapy (fluindione), started at day 2. The first 219 patients were randomized in a prospective study to determine the effect of a low dose of aspirin (200 mg) in the postoperative period (8). The other 461 patients did not receive any antiplatelet therapy.
Transesophageal echocardiography was performed in the early postoperative period in all patients, using a Hewlett-Packard Sonos 2500 or 5000 (Hewlett-Packard Imaging System, Andover, Massachusetts) and Acuson Sequoia (Acuson, Mountain View, California) connected to a multiplane 5-7 Mhz probe. All TEE were recorded on videotapes and reexamined twice. We carefully examined the atrial and ventricular side of the mitral prosthesis, looking for leaflet dysfunction, mitral regurgitation, spontaneous echocardiography contrast (SEC), strands, and thrombi. A strand was defined as a linear, mobile, and thin (<1 mm) echo attached to the atrial side of the prosthesis. A thrombus was defined as an abnormal well circumscribed mobile or immobile mass usually found in the vicinity of the mitral valve annulus. All characteristics of the thrombi were analyzed to specify size, form, localization, and mobility.
In accordance with the report of Guéret et al. (3), thrombi were defined as small when their maximal length was <5 mm (group A) and large when the length was 5 mm (group B).
Therapeutic approach and survey of thrombus
The therapy of the thrombi was left to the discretion of each surgical team and was analyzed retrospectively. Clinical follow-up, according to published recommendations (9), included total and cardiac mortality, along with morbid events: symptomatic embolic event (stroke, transient ischemic attack [TIA] of less than 24 h duration, peripheral embolism), prosthetic obstruction, and treatment complications such as hemorrhage. Transesophageal echocardiography monitoring was left to the discretion of the surgical and medical teams and was carried out in about half the cohort.
Early clinical course (one month) was studied during the in-hospital period. Long-term follow-up was conducted by telephone or letter contact to the consulting physicians at 34 ± 22 months since implantation.
All data are expressed as mean ± 1 SD. Qualitative values were compared by Fischer exact test. Unpaired Student ttest was used to the quantitative variables analysis. The p value indicating statistical significance was set at 0.05.
Early systematic postoperative TEE (day 9) was performed in 680 patients from June 1994 to December 2000, revealing in 64 patients the presence of one or more abnormal masses on the prosthesis interpreted as local thrombus (9.4%).
The cohort comprised 33 women and 31 men with a mean age of 64 ± 9 years (range, 39 to 82 years) (Table 1).
At the time of the TEE, all patients were treated by oral anticoagulation (fluindione), and 31 of them (47%) were still treated by subcutaneous or intravenous heparin when INR targets had not been reached.
The bileaflet valve was a St. Jude Medical prosthesis (St. Paul, Minnesota) in 38 cases, ATS (AB Medical, Minneapolis, Minnesota) for 14 patients, SORIN (Saluggia, Italy), and MIRA (Edwards Lifesciences, Mairepas, France) in six cases.
On day 9, only two patients were symptomatic: one presented with nonfatal stroke and the other with congestive heart failure (New York Heart Association class 4). However, these two patients presented a NOT on the prosthesis. Sixty-two others patients were either asymptomatic or had minor aspecific symptoms.
In two asymptomatic patients, transthoracic echocardiography (TTE) and TEE detected early OT, measuring more than 10 mm, and both were treated by redo-surgery (one mechanical valve, one bioprosthetic valve).
In the other 62 patients, transthoracic examination suggested normal function of the bileaflet prosthesis (no significant dysfunction of the poppet, no increase in transvalvular mean gradient). In many of them, TEE showed strands or spontaneous contrast echoes, but these phenomena were not considered as pathologic in our study.
Echographic characteristics of thrombi
Maximum diameter of thrombus was less than 5 mm in 29 cases (45.3%) (Figs. 1 and 2). ⇓Thrombus size was from 5 to 10 mm in 28 patients (43.7%), and thrombi of more than 10 mm were detected in seven patients (11%) including the two initial obstructive ones.
Sixty-two thrombi were attached to the atrial side of the prosthesis (on the sewing ring or on the poppet), but we also found four thrombi on the ventricular side of the annulus.
The abnormal mass was localized and round in 40 cases, and circumferential in 24 patients.
Thirty-six thrombi (56%) were sessile and immobile; 28 (44%) were pedunculated and slightly mobile.
Two predisposing factors of early thrombosis were analyzed: atrial fibrillation was found in 26 patients of 64 (41%) before or immediately after surgery. Sixteen patients (25%) were operated with an impaired left ventricle ejection fraction (<50%), and the association of both factors was observed in eight patients (12%). Three patients were implanted with a pacemaker in the postoperative period, and the anticoagulant therapy was, therefore, discontinued for a few hours.
Immediate therapy after the discovery of early postoperative thrombus was left to the discretion of each surgical team: the decision for reintervention was based on clinical and echographic data. The two early obstructive thrombi were immediately reoperated because of the major risk of complication; medical treatment was readjusted in all of the remaining cases of early NOT. Intravenous or subcutaneous heparin was reinitiated in 33 patients (53%) monitored by the aPTT. A simple readjustment and optimization of the oral anticoagulation (vitamin K antagonist therapy) with a target INR >3 was proposed for 19 patients (31%). Aspirin was added in 13 cases and nonsteroid anti-inflammatory therapy in seven patients.
Clinical course of patients with early thrombi
Immediate (<1 month) and long-term (=1 year) outcomes were analyzed in the cohort of 64 patients exhibiting an early thrombus after MVR (Table 1). The mean duration of follow-up was 34 ± 22 months and was obtained for the majority of patients, with only two lost to follow-up during the first year.
Early course (one month)
In group A (small thrombus, n = 29), we only observed one early (first month of follow-up) complicated course (3.4%): one patient had a TIA (defined as a neurologic syndrome of less than 24 h duration) and died of heart failure at postoperative day 15 (Patient 8).
Among the 35 patients of group B (large thrombus, n = 35), the early course seemed to be more serious. Two early obstructive thrombi were immediately reoperated, which is recognized as a complicated course (Patient 42 and Patient 63); in the other 33 patients with large NOT, six presented severe events: three strokes, defined as a sudden focal brain dysfunction lasting more than 24 h (Patients 2, 7, and 53), and two secondary enlargements, finally becoming obstructive in the first month despite anticoagulant therapy, and requiring redo-surgery (Patients 21 and 24); one of these patient died of multiple organ failure during the first month. In this cohort, we also observed another early fatal event: a patient with two NOT measuring 5 mm died as a result of gastrointestinal bleeding (Patient 13). In total, eight of the 35 patients with large thrombi presented an early (first month) complicated course (22%).
Long-term course (more than one year)
Over the whole follow-up period (mean duration: 34 ± 22 months), we observed two other complications in group A: two late TIA (Patient 19 in the second and Patient 40 in the fourth month). In group B, a trend for more serious complications was noted: two late nonfatal strokes (Patients 9 and 52) and three other deaths in this cohort. One patient with an early stroke died six months later (Patient 2), and another of heart failure at month 6 (Patient 41). One of the two patients with early OT, despite redo-surgery, presented a fatal course 12 weeks after MMVR (Patient 42).
It is of note that a patient with an early stroke had another neurological ischemic event five years later. In this patient, a small thrombus was seen on the valve, but this event was not recorded as a complication because of the disappearance of the first thrombus on TEE examination five months after implantation. Total mortality was estimated at 9.4% (six patients) in the long-term survey (three early and three late deaths), and five of these six patients were in group B.
Finally, in group A, the follow-up revealed a complicated course for 3.4% of patients in the first month and for 10% over the total duration of the study. In group B, complications occurred in 22% of patients in the early period and in 31% over the long-term survey. Complications were significantly more frequent during the early (p = 0.027) and long-term (p = 0.04) follow-up periods in the group of patients with large thrombi (group B).
From another viewpoint, we could split this large cohort into two different subsets: the first of 54 patients with an early (<1 month) uneventful course (in this group, mean size of thrombus = 5.56 ± 3.3 mm). The second group comprised 10 patients presenting an early complication after diagnosis of early thrombosis, whose mean lengths (8.95 ± 5.3 mm) were significantly higher than those of the first group (p = 0.0009). A significant difference was also noted between these two subsets in the long-term follow-up. We concluded that the early and long-term course appeared to depend on the initial size of the thrombus with a higher incidence of events (principally embolic accidents) for patients with large, often asymptomatic thrombi. These results are listed in Table 2and Figure 3and would indicate a revised therapeutic approach for patients with large thrombi.
There are numerous reports of a high incidence of postoperative NOT after MMVR (1–5,7). Our study represents the largest population to date of patients benefiting from systematic TEE after surgery. We showed that about 10% of these patients presented an abnormal mass usually adhering to the atrial face of the prosthesis, considered as a thrombus. It should be noted that all the patients except two examined in this cohort of 680 presented few or no symptoms on the day of early TEE.
Systematic TEE after MMVR
Diagnosis of mechanical valve thrombosis is usually based on TTE and cinefluoroscopy; on TTE, thrombosis can be detected from an increase in the transvalvular mean pressure gradient measured by Doppler. A reduction in mobility or even immobility of one of the poppets may also be evidenced. However, an increase in this gradient is commonly seen immediately after surgery due to the altered hemodynamic conditions, the presence of anemia or tachycardia, even when the prosthesis works perfectly. On the other hand, normal transthoracic echocardiogram examination and fluoroscopy may not exclude NOT. Thus, TTE and cinefluoroscopy seem to be attractive alternatives (5). However, TEE is recognized as having superior sensitivity and specificity for the diagnosis of mitral prosthetic thrombi because of echocardiography artefacts in the left atrium on TTE examination (10–14). Many studies have demonstrated the interest of TEE after mechanical MVR for the detection of thrombi (7–9). In our experience on 680 patients, we observed 64 thrombi (9.4%), most of them nonobstructive. Our patients (except two) presented few symptoms indicative of thrombotic phenomena, and their valves were considered to be functioning normally based on the postoperative Doppler parameters from systematic TTE. This incidence had been demonstrated in many prospective studies of systematic TEE after MVR. However, in these asymptomatic patients, we observed a higher incidence of early and late complications, such as systemic embolic events (TIA, stroke) or obstruction with heart failure. We found a relationship between the initial size of the thrombus and the incidence of complications. These observations would indicate the value for a thorough search for early thrombosis after MMVR, as a guide to postoperative management. We, therefore, recommend, as mentioned in French guidelines for the management of patients with valvular heart diseases (15), systematic postoperative TEE after MMVR.
Predisposing factors for thrombosis
Many authors have looked for hemodynamic or echographic factors predictive of early thrombi after mitral surgery. Dadez et al. (2)showed that SEC in the left atrium as detected by TEE was associated with abnormal small echoes on the atrial side of the prosthesis, which were responsible for the increase in embolic events. This notion was supported by Bonnefoy et al. (4)in a study of systematic TEE 24 h after MVR. Özkan (16)described three independent predictors: SEC, atrial fibrillation, and preoperative thrombus. However, Malergue et al. (1)failed to evidence any statistical difference for incidence of atrial fibrillation, SEC, or left ventricle dysfunction. Guéret (3)concluded that early thrombosis after mitral prosthesis was associated with unpaired systolic fractional shortening and atrial parameters. According to Iung (6), small abnormal echos after MMVR are associated with conditions predisposing to SEC with high dose protamine administration during surgery compounded by inadequate postoperative anticoagulation.
In a previous study in our center (7), the low level of anticoagulation in the immediate postoperative period emerged as the principal cause of early thrombosis. Indeed, the formation of a thrombus can be explained by an imbalance in the parameters of Virchow (17): cardiac surgery, especially mechanical MVR, and extracorporeal circulation damages native tissues, while artificial surfaces induce variations in local rheology with stagnation of blood and reduction in flow velocity. A hypercoagulable state is often observed with platelet and leucocyte activation, with coagulation factors and fibrinogen maintaining a vicious circle. All these parameters are affected by poor anticoagulation, which is thought to represent the most important risk factor (3,6,7).
Management of thrombus after mechanical MVR is still matter of debate (18). For obstructive thrombi, the major risk of complications indicates an early aggressive approach according to American College of Cardiology/American Heart Association guidelines (19). Fibrinolysis is contraindicated in the first 14 postoperative days due to risk of major bleeding (hemopericardium), but may be indicated at later times (18,20–26). Guéret (3)recommends systematic surgical treatment for obstructive or large and threatening thrombi, or failure of medical therapy (complications or thrombi growing under anticoagulant therapy). Although redo-surgery carries a major risk of mortality (7), we applied this strategy to the two initial and two secondary OT, with long-term success in three of these patients.
The problem of management of NOT is more acute (24). The evolution of these thrombi seems to depend on the initial size as found in the present study. Several authors have shown the importance of anticoagulation in the immediate postoperative period. We also recommend the reinitialization of intravenous continue heparin monitored by aPTT until correct adjustment of oral anticoagulation.
The value of aspirin is still unclear. Laffort et al. (8)showed that aspirin added to conventional oral anticoagulation decreased the incidence of postoperative thrombi, but long-term prescription increased the incidence of major bleeding (gastrointestinal hemorrhage). We propose an association of anticoagulation and aspirin with gastric protection for large and mobile NOT; this group of patients will probably require TEE survey until disappearance of the thrombi. However, the incidence of potential thromboembolic events in the subset with large thrombi might justify a more aggressive approach.
The principal limitation of the present study was the absence of randomization in the therapeutic strategies. Each surgical team chose how to manage their patients after diagnosis of the thrombosis, and we did not influence their decisions. Apart from OT for which redo-surgery seems essential, we could not draw any conclusions about optimal treatment of such early thrombi.
At the beginning of our study, a systematic TEE was performed at the fifth month (26 patients), but we did not extend this to the whole population, and so we lack information on the echographic outcome of all thrombi, especially in the patients with uneventful course.
This study represents the largest cohort of patients benefiting from TEE after MMVR and confirms the high incidence (about 10%) of early postoperative, often asymptomatic, and NOT.
The course is usually benign for small thrombi (<5 mm) when anticoagulant therapy is adapted. Patients presenting large NOT develop more complications and require a more aggressive approach: reintroduction of heparin until oral anticoagulation is well adjusted, administration of aspirin, and, in the most threatening large thrombi, redo-surgery.
☆ Robert A. O'Rourke, MD, acted as the guest editor for this paper.
- activated partial thromboplastin time
- international normalized ratio
- mechanical mitral valve replacement
- nonobstructive thrombus
- obstructive thrombus
- spontaneous echocardiographic contrast
- transesophageal echocardiography
- transient ischemic accident
- transthoracic echocardiography
- Received March 13, 2003.
- Revision received August 19, 2003.
- Accepted September 23, 2003.
- American College of Cardiology Foundation
- Guéret P.,
- Vignon P.,
- Fournier P.,
- et al.
- Laffort P.,
- Roudaut R.,
- Roques X.,
- et al.
- Roudaut R.,
- Labbé T.,
- Marcaggi X.,
- et al.
- Habib G.,
- Cornen A.,
- Mesana T.,
- et al.
- Barbetseas J.,
- Nagueh S.,
- Pitsavos C.,
- et al.
- ↵Virchow R. Phlogose und thrombose in gefessystem. Gessammelte Abhandlungen zur Wishenschaftlichen Medicin. Frankfurtam-Main, Germany: Meidiger Sohn and Company, 1856:458–63.
- Alpert J.S.
- Bonow R.O.,
- Carabello B.,
- de Leon A.C.,
- et al.
- Roudaut R.,
- Lafitte S.,
- Lorient-Roudaut M.F.,
- et al.
- Özkan M.,
- Kaymaz C.,
- Kirma C.,
- et al.
- Shapira Y.,
- Herz I.,
- Vaturi M.,
- et al.
- Lengyel M.,
- Fuster V.,
- Keltai M.,
- et al.
- Silber H.,
- Khan S.,
- Matloff J.,
- et al.