Ann Thorac Surg 2002;73:1479-1483
© 2002 The Society of Thoracic Surgeons
Original article: cardiovascular
Catheter pericardiocentesis for delayed tamponade after cardiac valve operation
Abeel A. Mangi, MDa,
Igor F. Palacios, MDb,
David F. Torchiana, MD*a
a Division of Cardiac Surgery, Massachusetts General Hospital, Harvard Medical School, Boston, Massachusetts, USA
b Division of Cardiology, Massachusetts General Hospital, Harvard Medical School, Boston, Massachusetts, USA
Accepted for publication February 2, 2002.
* Address reprint requests to Dr Torchiana, Division of Cardiac Surgery, Bullfinch 112, Massachusetts General Hospital, 55 Fruit St, Boston, MA, USA 02114
e-mail: dtorchiana{at}partners.org
 |
Abstract
|
|---|
Background. Late tamponade is a rare cause of morbidity and mortality after cardiac valve operation. We describe our recent experience with this entity.
Methods. This is a single institution, procedure-matched, retrospective review of patients undergoing pericardiocentesis more than 7 days after cardiac operation, during a 7-year period.
Results. Pericardiocentesis for delayed tamponade was performed in 43 of 9,612 patients. Although isolated valve operation accounted for 17% of all patients overall, 76% of patients undergoing pericardiocentesis (33 of 43) had undergone isolated valve operation. The average age in this group was 58 years, compared to an average of 68 years in all patients. Patients presented with tamponade an average of 18 days after operation. Positive predictors included elevated prothrombin time on presentation. Of the patient cohort 75% presented with dyspnea, 61% with inability to diurese, and 61% with hypotension. Echocardiography detected effusions in all patients, but specific echocardiographic signs of tamponade were present in only 30%. Of the patients, 97% were successfully treated by pericardiocentesis. All were safely restarted on warfarin. One patient required pericardial window.
Conclusions. Delayed cardiac tamponade is more common after isolated valve operation, as opposed to coronary artery bypass grafting and valve/coronary artery bypass grafting. It tends to occur in the third postoperative week in younger patients who are aggressively anticoagulated. Pericardiocentesis with catheter placement is highly effective, and patients can be reanticoagulated safely. This series underestimates the incidence of late tamponade, as some patients may present to outside facilities. The diagnosis is aided by a high degree of suspicion.
|
Introduction
|
|---|
Cardiac tamponade is a potentially lethal complication after open heart operation, and occurs in 0.5% to 6% of cases [1, 2]. Tamponade can either occur in the early or late postoperative period. Early postoperative tamponade is more common, and is easier to recognize because of close hemodynamic monitoring in the intensive care unit or recovery room. Late postoperative tamponade was first documented in 1968 by Prewitt and colleagues [3] and by definition, occurs more than 7 days after cardiac operation. It has been reported to occur up to 6 months after cardiac operation [4]. With the majority of cardiac surgical patients now being discharged from the hospital by the fifth postoperative day, it is increasingly likely that the initial encounter of the patient with late tamponade will not be with the cardiac surgeon. The varied presentation of this entity makes diagnosis challenging, with the result that mortality rates as high as 18% have been reported [5]. The younger age of this patient population has been documented [6]. Increased awareness of this entity among clinicians can lead to early recognition, appropriate management, and the prevention of morbidity and mortality in these young and often otherwise healthy patients.
This retrospective procedure-matched study is the largest from a single hospital. Our aim is to document preoperative risk factors, postoperative symptoms and signs of late tamponade, and outline appropriate strategies for management of this entity.
|
Material and methods
|
|---|
Patient selection and chart review
Nine thousand six hundred twelve patients underwent cardiac valve operation or bypass grafting by eight surgeons at the Massachusetts General Hospital during an 86-month period from 1992 to 1999. Of the patients, 72.9% underwent coronary artery bypass grafting (CABG) alone, 17.6% underwent valve replacement/repair alone, and 9.5% underwent CABG combined with valve replacement/repair. After approval of the institutional review board, the records of the catheterization laboratory database were reviewed and referenced against the surgical registry. Forty-three patients underwent pericardiocentesis more than 7 days after cardiac operation. Thirty-three (76%) of these patients had undergone valve grafting alone. Thirty-three procedure-matched patients were randomly chosen as a control group for this cohort. Follow-up data were obtained by review of office charts, emergency room charts, and hospital records.
Pericardiocentesis technique
Pericardiocentesis was performed in the catheterization laboratory through a subxiphoid approach under electrocardiographic and fluoroscopic guidance. Arterial and right heart catheterization monitoring were frequently used. After administration of 1% lidocaine to the skin and the deeper tissues of the left xiphocostal area, the pericardial needle was connected to an electrocardiogram V lead. The needle was advanced from the left of the subxiphoid area aiming toward the left shoulder. A discrete pop was occasionally felt as the needle entered the pericardial space. Sometimes ST segment elevation was seen on the V lead tracing when the needle touched the epicardium. Under this circumstance, the needle was retracted slightly until ST segment elevation disappeared. Once the pericardial space was entered, a guidewire was introduced into the pericardial space through the needle. The needle was removed and a catheter was inserted into the pericardial sac over the guidewire. A pigtail-shaped drainage catheter with an end and multiple side holes was used. Intrapericardial pressure was measured by attaching a pressure transducer system to the intrapericardial catheter. Pericardial fluid was then removed. Aspiration of fluid was continued until there was clear clinical and hemodynamic improvement. The catheter was frequently left in place to monitor pericardial fluid drainage. The catheter was secured to the skin with 4-0 silk sutures, covered with a sterile dressing and the patient was started on prophylactic antibiotics. The pericardial space was drained every 8 hours and the catheter flushed with heparinized solution and in general removed within the next 24 to 72 hours.
Statistical methods
The world literature was reviewed (MEDLINE search, 1966 to 2001) and a secondary review of referenced works was performed. Statistical analysis was performed using Statview for Windows (v. 5.0.1, SAS Institute, Cary, NC). Differences between groups were demonstrated using the 
2 test, and the Fisher’s exact test where appropriate. A p value of less than 0.05 was necessary to achieve statistical significance.
|
Results
|
|---|
Patient and procedure characteristics
Delayed cardiac tamponade is significantly more common in patients undergoing valve operation alone, as compared to CABG alone or concomitant CABG/valve grafting. At our hospital, only 1 of 917 patients (0.001%) undergoing CABG/valve grafting, and 9 of 728 patients (0.001%) undergoing CABG alone presented for pericardiocentesis with delayed cardiac tamponade using the criteria set forth in this study. In contrast, 33 (the cohort group) of 1,667, or almost 2% of patients undergoing valve operation alone, had pericardiocentesis for late postoperative cardiac tamponade, using the same criteria. This fourfold order of magnitude difference was statistically significant (p < 0.005).
Characteristics of the 33 patients in the cohort group were generally comparable to those in a procedure-matched control group (Table 1).
Patients in the cohort group were younger than those in the control group (58 years versus 68 years; p < 0.05). Patient sex, preoperative medications such as aspirin and warfarin, tendencies toward bleeding diatheses (such as hemophilia), and valve type used were comparable in the two groups. The immediate postoperative characteristics in both groups were also comparable, with similar numbers requiring reexploration for bleeding, similar numbers demonstrating atrial dysrhythmias, similar doses of diuretics, and warfarin used in the postoperative period. At our hospital, the targeted international normalized ratio is 3.5 for mechanical mitral valve replacement and 2.5 for aortic valve replacement, bioprosthetic mitral valve replacement, and mitral valve repair. Warfarin is started on the first postoperative day in stable patients. All patients in this study were on warfarin at the time of presentation. Postoperatively, heparin is used sparingly—usually only in patients who are believed to be at increased risk of thromboembolic complications, in whom the international normalized ratio is subtherapeutic on the fifth postoperative day. All patients receive anterior and posterior mediastinal thoracostomy tubes. In the absence of air leaks, criteria for removal of thoracostomy tubes include drainage of less than 10 mL per tube per hour. Approximately 80% of thoracostomy tubes meet these criteria, and are removed on postoperative day 1. Surgeon preference dictates the use of concomitant aspirin and warfarin, which enabled us to demonstrate a significant correlation between their combined use, and delayed postoperative cardiac tamponade (Table 2).
Clinical presentation
Patients presented with late postoperative cardiac tamponade approximately 2.5 weeks (18.6 ± 12.1 [standard deviation] days) after operation. Symptoms and signs were documented at the time of the initial patient encounter, which was most often in the physician’s office or emergency ward. The most common symptom was dyspnea (25 patients in the cohort group versus 9 in the control group; p < 0.05). The most common signs were hypotension (systolic blood pressure less than 90 mm Hg) and an inability to diurese (assessed by an increase in body weight despite increasing doses of diuretics). Both these signs reached statistical significance when compared to the control group (Table 2). The set of symptoms and signs commonly associated with pericardial effusion such as muffled heart tones, friction rub, and pain on sitting up and leaning forward were commonly present in both groups. A pulsus paradoxus was documented in only one-third of patients, and was meaningful (ie, an inspiratory decrease in systolic pressure of 15 mm Hg or greater) in all of these patients. This sign was not routinely documented in the control group. Echocardiogram demonstrated effusions in all 33 patients, but specific echocardiographic signs of tamponade such as right atrial collapse or septal deviation were present in only one-third of patients examined. A highly significant finding was an elevated prothrombin time in the cohort group (21.4 ± 6.9 seconds) when compared to the control group (17.3 ± 2.3 seconds; p < 0.005) (Fig 1).
Management and outcome
Twenty-one patients were taken directly to the cardiac catheterization laboratory. These were patients with hemodynamic compromise or in whom the echocardiogram demonstrated right atrial collapse. The remaining patients were hemodynamically stable, and did not have specific echocardiographic findings of cardiac tamponade. They were treated empirically with oral indomethacin, and often with aggressive attempts at diuresis. They failed to improve and ultimately went on to pericardiocentesis. The average opening intrapericardial pressure was 16.6 mm Hg (standard deviation 6.9 mm Hg), and decreased to an average of 2 mm Hg after drainage of 120 mL to 1,700 mL (mean 788.5 ± 384 mL) of pericardial fluid. Seventy-nine percent of patients had an indwelling pigtail catheter left in place, which was removed 2 to 5 days after the procedure. All patients were able to resume anticoagulant therapy or antiplatelet therapy. Only 1 patient required a repeat pericardiocentesis, and after reaccumulating a second time, required a pericardial window. Patients did not undergo routine postprocedure echocardiogram. All patients survived the procedure without complications such as infection or pneumothorax, and all remained free of disease at more than 1 year of follow-up (Table 3).
|
Comment
|
|---|
Few would dispute the necessity for anticoagulation after cardiac valve operation in most patients. Judicious administration of warfarin and careful monitoring of the prothrombin time are necessary to strike the balance between the prevention of thromboembolic complications and the avoidance of bleeding [7, 8]. Multiple investigators have asserted that excessive anticoagulation in the postoperative period is responsible for late postoperative tamponade [9–14]. In this study, we found that patients with late postoperative tamponade have statistically significant higher prothrombin times at presentation than matched control patients. In addition, these patients are more likely to be on both aspirin and warfarin in the postoperative period. This theory is also supported by the significantly lower age of the cohort group, suggesting that mechanical valve replacement (more commonly used in younger patients, with a more aggressive anticoagulation regimen) may render patients more susceptible to late tamponade than bioprosthetic replacement, or valve repair. One explanation that has been advanced to support this hypothesis suggests that patients undergoing mechanical valve replacement are somehow rendered more intrinsically sensitive to warfarin [15]. This line of thinking finds support in studies of older mechanical valves such as the Beall valve (Coratomic, San Diego, CA), which demonstrated abnormal coagulation profiles with advanced wear [16], issues that may not be as pertinent with newer generation mechanical valves. However, recent studies suggest that patients undergoing aortic valve replacement are, to the contrary, hypercoagulable, with elevated levels of thrombin-antithrombin III complex, modified antithrombin and protein S levels [17]. Finally, in this series, 67% of patients underwent mechanical valve replacement in the cohort group, and 65% underwent mechanical valve replacement in the control group. This similarity argues against the explanation that mechanical valve replacement or international normalized ratio target is sufficient to explain a tendency toward late tamponade. A second theory is that hepatic dysfunction secondary to tamponade may result in an accentuated response to warfarin therapy, resulting in a higher international normalized ratio and eventually, in more bleeding [18]. The inciting event behind the origin of tamponade is not satisfactorily explained by this theory, but is presumably related to postoperative pericarditis and inflammatory changes.
One of the difficulties in safely administering warfarin is the number of interactions it has with other drugs. Prospective randomized trials have demonstrated no obvious difference in cephalosporins for prophylaxis in cardiac operations [19, 20], with vancomycin proving to be slightly superior [21]. Cephalosporins and vancomycin are known to augment prothrombin times by approximately 50% [22] when given individually. Whether or not this effect is further augmented by simultaneous administration has not been studied. At our hospital, perioperative antibiotic coverage is instituted on call to the operating room, and continued for 48 hours postoperatively. Warfarin therapy is typically instituted 24 hours postoperatively. Choice of antibiotic coverage is surgeon dependent. Nineteen patients in the cohort group (compared with 10 in the control group; p > 0.05) were treated with both vancomycin and cefazolin, suggesting that the two drugs may act synergistically to potentiate the effect of warfarin very early in the postoperative course. Should bleeding occur at this time, it can be potentially aggravated by removal of thoracostomy drains and pacing wires. In addition, these effects of hepatic dysfunction can serve to further augment prothrombin times, thereby resulting in grossly elevated prothrombin times at presentation. This may explain why cases of late postoperative tamponade do not appear to be increasing despite the widespread use of platelet aggregation inhibitors such as clopidogrel (Plavix) and epfitibadite (Integrilin), and why patients are able to safely resume anticoagulation after pericardiocentesis.
Late cardiac tamponade is not seen as often after CABG or concomitant valve/CABG. One explanation is that warfarin is not routinely used after CABG. In the case of concomitant valve/CABG, an alternative explanation is that opening the left pleural space to harvest the left internal mammary artery may allow drainage of mediastinal fluid into the pleural space, thereby preventing tamponade. At our hospital, the pleural space is routinely widely opened during harvest of the left internal mammary artery.
The major limitation of this study is methodological. Screening just the cardiac catheterization database for patients undergoing postoperative pericardiocentesis excludes patients who may have been treated by other means at our hospital, as well as patients who presented to, and were treated at other hospitals. This study, therefore, underestimates the morbidity and mortality associated with late cardiac tamponade as well as its incidence. In addition, it carries the biases inherent in retrospective analysis, and is limited by the relatively small sample size.
In summary, late postoperative cardiac tamponade is an uncommon, and potentially lethal condition. It may have its origins in pharmacologic interactions in the perioperative period, and is potentially preventable by cautious postoperative warfarin therapy, and by opening of the pleural space at the time of valve operation. Failing these measures, the patient with late postoperative tamponade presents with a clinical syndrome that may include malaise, failure to diurese, prerenal azotemia, poor exercise tolerance, dyspnea, hypotension, and a pulsus paradoxus. The diagnosis is clinical and should be suspected in any cardiac surgical patient on coumadin who is failing to thrive early after discharge. Echocardiography may fail to demonstrate specific signs of tamponade beyond demonstrating the presence of an effusion. As some patients are hemodynamically compromised, aggressive diuresis is unwise and may aggravate hypotension or hypoperfusion. Pericardiocentesis is the treatment of choice, and should be performed based on the clinical picture, not on echocardiographic criteria alone. This is usually the only treatment required, after which all patients can safely resume anticoagulant therapy.
|
Acknowledgments
|
|---|
We thank Sandra L. DeBronkart, RN, for her invaluable assistance in compiling the data used in this study.
|
References
|
|---|
-
Craddock D.R., Logan A., Fadali A. Reoperation for hemorrhage following cardiopulmonary bypass. Br J Surg 1968;55:17-20.[Medline]
-
Stevensen L.W., Child J.S., Laks H., Kern L. Incidence and significance of early pericardial effusions after cardiac surgery. Am J Cardiol 1984;54:848-851.[Medline]
-
Prewitt T.A., Rackley C.E., Wilcox B.R., Scatliff J.H., Young D.T. Cardiac tamponade as a late complication of open heart surgery. Am Heart J 1968;76:139-141.[Medline]
-
Ofori-Karkye S.K., Tyberg T.I., Geha A.S., Hammond G.L., Cohen L.W., Langou R.A. Late cardiac tamponade after open heart surgery: incidence, role of anticoagulants in its pathogenesis and its relationship to postpericardiectomy syndrome. Circulation 1981;63:1323-1328.[Free Full Text]
-
Garcia J.M., Reyes E., Cheanvecchai C., Effler D.B. Delayed cardiac tamponade following open heart surgery. Cleveland Clin Quart 1974;41:103-108.
-
Premaratne S., Zhang W., Arakaki H.Y., Nakasone C.K., McNamara J.J. Effect of anticoagulants of the incidence of late pericardial tamponade following open heart surgery: the Hawaii experience. J Cardiovasc Surg (Torino) 1997;38:107-111.[Medline]
-
Horstkotte D., Schulte H.D., Bircks W., Strauer B.E. Lower intensity anticoagulation therapy results in lower complication rates with the St. Jude Medical prosthesis. J Thorac Cardiovasc Surg 1994;107:1136-1145.[Abstract/Free Full Text]
-
Cannegieter S.C., Rosendaal F.R., Briet E. Thromboembolic and bleeding complications in patients with mechanical heart valve prostheses. Circulation 1994;89:635-641.[Abstract/Free Full Text]
-
Hsin M., Speight J., Cooper G.J. Anticoagulant control in the first three months after mechanical heart valve replacement. Int J Clin Pract 1997;51:427-429.[Medline]
-
Bortolotti U., Livi U., Frugoni C., et al. Delayed cardiac tamponade following open heart surgery. Analysis of 12 patients. Thorac Cardiovasc Surg 1981;29:233-236.[Medline]
-
Bashi V.V., Ravikumar E., John S. Delayed tamponade following open heart surgery: experience with twelve patients. Indian Heart J 1987;39:206-211.[Medline]
-
Ellison L.H., Kirsh M.M. Delayed mediastinal tamponade after open heart surgery. Chest 1974;65:64-66.[Medline]
-
Sahni J., Ivert T., Herzfeld I., Brodin L.A. Late cardiac tamponade after open heart surgery. Scan J Thorac Cardiovasc Surg 1991;25:63-68.[Medline]
-
Solem J.O., Kugelberg J., Stahl E., Olin C. Late cardiac tamponade following open heart surgery. Diagnosis and treatment. Scan J Thorac Cardiovasc Surg 1986;20:129-131.[Medline]
-
Rose J.P., Rihn T.L., Long S.F. Warfarin sensitivity after mechanical heart valve replacement. Pharmacotherapy 1998;18:856-859.[Medline]
-
Clark R.E., Pavlovic T.A., Knight B.E., et al. Quantification of wear, hemolysis and coagulation deficits in patients with Beall mitral valves. Circulation 1977;56(suppl 2):139-143.[Free Full Text]
-
Welters I., Menges T., Ballesteros N., et al. Thrombin generation and activation of the thrombomodulin protein C system in open heart surgery depend on the underlying cardiac disease. Thromb Res 1998;92:1-9.[Medline]
-
Wong P.S., Pugsely W.B. Raised international normalized ratio (INR): is it a cause or an effect of late cardiac tamponade?. Br Heart J 1992;68:212-213.[Abstract/Free Full Text]
-
Townsend T.R., Reitz B.A., Bilker W.B., Bartlett J.G. Clinical trial of cefamandole, cefazolin, and cefuroxime for antibiotic prophylaxis in cardiac operations. J Thorac Cardiovasc Surg 1993 Oct;106(4):664-670.[Abstract]
-
Gentry L.O., Zeluff B.J., Cooley D.A. Antibiotic prophylaxis in open-heart surgery. A comparison of cefamandole, cefuroxime, and cefazolin. Ann Thorac Surg 1988;46:167-171.[Abstract/Free Full Text]
-
Maki D.G., Bohn M.J., Stolz S.M., Kroncke G.M., Acher C.W., Myerowitz P.D. Comparative study of cefazolin, cefamandole, and vancomycin for surgical prophylaxis in cardiac and vascular operations. A double-blind randomized trial. J Thorac Cardiovasc Surg 1992;104:1423-1434.[Abstract]
-
Angaran D.M., Dias V.C., Arom K.V., et al. The comparative influence of prophylactic antibiotics on the prothrombin response to warfarin in the postoperative prosthetic cardiac valve patient. Cefamandole, cefazolin, vancomycin. Ann Surg 1987;206:155-161.[Medline]
This article has been cited by other articles:
|
|
|
|
 
A. A. Mangi and D. F. Torchiana
Pericardial Disease
,
January 1, 2008;
3(2008):
1465 - 1478.
[Full Text]
|
|
|
|
|
|
|
A. Kulik, F. D. Rubens, P. S. Wells, C. Kearon, T. G. Mesana, J. van Berkom, and B.-K. Lam
Early Postoperative Anticoagulation After Mechanical Valve Replacement: A Systematic Review
Ann. Thorac. Surg.,
February 1, 2006;
81(2):
770 - 781.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
A. A. Mangi and D. F. Torchiana
Pericardial Disease
,
January 1, 2003;
2(2003):
1359 - 1372.
[Full Text]
|
|
|
Top
Abstract
Introduction
