Frequent Occurrence of Occult Pulmonary Embolism From Venous Sheaths During Endomyocardial Biopsy

Transcription

1 JACC Vol. 19, No Frequent Occurrence of Occult Pulmonary Embolism From Venous Sheaths During Endomyocardial Biopsy SUSAN K. KREHER, MD, FACC, VALERIE K. ULSTAD, MD, CANDACE D. DICK, MD, RUTH DEGROFF, RDMS, MARIA THERESA OLIVARI, MD, DAVID C. HOMANS, MD, FACC Minneapolis, Minnesota To determine the frequency of occult right heart thromboembolism during endomyocardial biopsy, 51 cardiac transplant recipients undergoing routine endomyocardial biopsy were studied echocardiographically. Patients were randomized to two groups. In Group 1, the venous sheath was flushed between each biopsy attempt; in Group 2, it was flushed only at the time of initial placement. Right heart thromboemboli were identified in 18 (35%) of 51 patients. Seventeen (94%) of these 18 patients were in Group 2. Patients requiring more than six biopsy attempts had a signifi- cantly higher incidence of embolism. Other variables such as antiplatelet therapy, operator experience and total time of the procedure did not correlate with occurrence of thrombus. All right heart emboli were asymptomatic. These data demonstrate a high incidence of occult pulmonary embolism during uncomplicated routine endomyocardial biopsy. Meticulous flushing of the introducer sheath significantly reduces the incidence of thrombus formation in intravenous sheaths. (J Am Coli CardioI1992;19:581-5) Central venous sheaths and catheters have become routine adjuncts to diagnostic and therapeutic regimens and are placed in patients by physicians with all levels of training. It has been observed (1) that long-term indwelling catheters and lines develop adherent thrombus that may embolize. Less is known about short-term venous access, the time required for thrombus to develop, the severity of embolic events and the best method for preventing thrombus formation. This study was designed to determine the incidence of right heart embolism from newly placed central venous sheaths during endomyocardial biopsy. Echocardiography was used to identify right heart thromboemboli. Although echocardiography has been described as a useful technique for guiding endomyocardial biopsy (2,3), its ability to detect right heart thromboemboli during endomyocardial biopsy has not been described. Methods Study patients. The study group consisted of 51 heart transplant recipients at the University of Minnesota undergoing routine rejection surveillance endomyocardial biopsy. Patients were informed of the intent of the study and signed informed consent. The study was approved by the Commit- From the Department of Medicine, Cardiovascular Division, University of Minnesota, Minneapolis, Minnesota. Manuscript received May I, 1991; revised manuscript received August 19, 1991, accepted August 27, Address for reprints: David C. Homans, MD, Cardiovascular Division, Box 58 UMHC, University of Minnesota, Minneapolis, Minnesota by the American College of Cardiology tee on the Use of Human Subjects in Research at the University of Minnesota. Procedure. With standard sterile technique, a 6P polypropylene sheath with side-port (Cordis catheter sheath introducer system) was percutaneously placed in the right internal jugular vein. The electrocardiogram (ECG) was monitored and the sheath was initially flushed with heparinized saline solution in all patients. Patients were then randomized to a "flush" or "no flush" protocol. In Group 1, the introducer sheath was flushed with heparinized saline solution between each biopsy attempt; in Group 2, it was not flushed after the initial placement. Sheaths were flushed by aspirating 3 to 5 ml of blood through the side-port and then vigorously infusing 5 to 1 ml of heparinized saline solution by hand. Fluid aspirated from the sheath was emptied from the syringe onto gauze and visually inspected for thrombus. Under fluoroscopic and echocardiographic guidance, four or five endocardial biopsy specimens were obtained by reo peated introduction of the biopsy forceps (Cordis biopsy forceps, 5AP) into the right ventricle through the introducer sheath (4,5). An echocardiogram in an apical, modified apical or subcostal view was recorded during each biopsy attempt. Echocardiograms were performed with use of a Hewlett Packard Sonos 5 cardiac ultrasound system or an Aloka 87 cardiac ultrasound system. The highest frequency transducer consistent with adequate ultrasound penetration was chosen on an individual basis. Echocardiographic imaging began shortly before the biopsy forceps was inserted into the sheath and was continued for the duration of each biopsy attempt. All sheaths were inspected for thrombus after removal from the patient. After the procedure, all patients /92/$5.

2 582 KREHER ET AL. lacc Vol. 19, No.3 Figure 1. Left, Right heart thromboembolus attached to the tricuspid valve (TV) (arrow). Asterisk denotes the right atrial (RA) suture line. Right, Diagrammatic portrayal of thrombus adherent to the tricuspid valve. LA = left atrium; MV = mitral valve; RV = right ventricle. were questioned regarding chest pain, dyspnea and other symptoms. Data acquisition. The number of biopsy attempts, total time the sheath was in place, use of antiplatelet or anticoagulant medications and level of training of the individual performing the biopsy were recorded for each procedure. Data analysis. Echocardiograms were reviewed independently by two observers. Discrepancies were resolved by a third independent echocardiographer, whose decision was final. Thromboemboli were defined as mobile, highly echogenic structures, generally serpiginous or oblong in shape, that could be seen on sequential frames to move through the right heart chambers (Fig. I to 3). These were significantly larger then the microbubbles that result from flushing a venous sheath. A biopsy attempt was defined as any insertion of the biopsy forceps into the sheath, whether or not myocardial tissue was obtained. Statistical analysis of the results was performed with use of chi-square analysis and the Fisher exact test. Multivariate analysis was then performed, using the variables with the most significance from the chi-square analysis. Results Occurrence of thromboemboli (Table 1). Fifty-one patients underwent endomyocardial biopsy with echocardiographic monitoring; 274 biopsy attempts were recorded and reviewed by two independent observers. In 96.4% of cases, Figure 2. Echocardiogram (left) and diagram (right) of right heart thrombus attached to the distal tip of the biopsy forceps (arrows). The biopsy forceps have been advanced well into the right atrium (RA), nearly to the plane of the tricuspid anulus. The thrombus, which appeared freely mobile on real time echocardiography, has crossed the plane of the tricuspid anulus and extends into the cavity of the right ventricle (RV). LV = left ventricle.

3 JACC Vol. 19, No.3 KREHER ET AL. 583 Figure 3. Sequential images (top) with diagrams (bottom) demonstrating the passage of the embolus through the right atrium (RA) and right ventricle (RV). Left panels, The thrombus can be initially seen in the mid-right atrium (arrow). Middle panels, The thrombus is approaching the plane of the tricuspid anulus (arrow). Right panels, The thrombus has advanced to the midright ventricle (arrow). the readers agreed on the interpretation; 1 biopsy attempts.6%) required review by a third echocardiographer. Only 1 of the 18 patients who had thromboemboli visualized during the biopsy was in Group 1; in that patient, a small embolus was noted on the 11th biopsy attempt. Thus, 17 (65%) of 26 patients in Group 2 had emboli (p <.1) versus 1(3.7%) of 27 in Group 1. No emboli were seen in 33 of the 51 of the patients; 24 (73%) of these were in Group 1and 9 (27%) were in Group 2. Variables affecting the incidence of thromboemboli. The number of biopsy attempts was also compared with the incidence of thromboembolism. When fewer than six attempts were required, only 1 of 4 patients had thrombi visualized (Fig. 4). Among patients with more than six '?... Table 1. Influence of Flushing and Number of Biopsy Attempts '-' S on Occurrence of Intrasheath Thrombus >3... Thrombus Thrombus % of Group ;:l 25-3 I " Present Absent With Thrombus 2-25 () Group P-. <6 attempts >6 attempts Group 2 i=: -5 <6 attempts ,... I aj >6 attempts Group 1 = routine flushing; Group 2 = no routine flushing. biopsy attempts, 8 of 11 had thrombi visualized (p <.1). There was no statistically significant relation between the duration of the procedure and the frequency of right heart embolism (Fig. 5). The incidence of embolism was not significantly affected by the level of operator experience or long-term use of aspirin or dipyridamole. In no case were thrombi associated with clinical findings of chest pain, shortness of breath, clinical instability, Figure 4. Stacked bar graph demonstrating the relation between right heart emboli and the duration of the biopsy procedure (in minutes). There was no statistical relation between the duration of the biopsy procedure and the incidence of right heart thromboemboli identified by echocardiography.... ;:l Q - Clot present o Clot absent Number of Patients

4 584 KREHER ET AL. lacc Vol. 19. No.3 March I. 1992:581-5 rn +J 25 I:l... 2 _ Clot Present +J til o Clot Absent fl., s.. 1.c 8 5 ;:; Z o >1 Number of Biopsy Attempts Figure 5. Stacked bar graph demonstrating the relation between the number of biopsy attempts and the incidence of right heart thromboemboli during endomyocardial biopsy. There was a statistically significant increase in the incidence of right heart thromboemboli in procedures requiring more than six biopsy attempts. changes in heart rate or rhythm or appearance of a QRS complex on the ECG. Thrombi aspirated from the venous sheath. Seven thrombi, four in Group 2 and three in Group 1, were aspirated from the venous sheath during biopsy. Right heart thromboemboli were identified echocardiographically in two of the four patients in Group 2but in none of the three patients in Group 1. All seven thrombi were small, generally measuring 2 to 3 mm in diameter and 7 to 14 mm in length. Discussion Occult right heart thromboembolism. This study demonstrates that occult pulmonary embolism is a common occurrence during endomyocardial biopsy and that meticulous flushing of the sheath significantly decreases the incidence of embolism. Furthermore, thromboemboli were more frequent when there were more than six biopsy attempts per patient. They were readily visualized with two-dimensional echocardiography and characteristically occurred when the bioptome first entered the right atrium. Occasionally, portions of thrombus adhering to the bioptome were visualized. These findings suggest that thrombus forms in or on the sheath between biopsy attempts and is dislodged when the bioptome is advanced through the sheath. Thrombi were also aspirated from the sheath between individual biopsy attempts and immediately after the sheath had been removed from the patient. Previous studies. An in vitro study of thrombus formation in angiographic catheters by Raininko and Riihela (6) used blood-filled catheters at 37 C, which had been previously filled with either physiologic saline solution, heparinized saline solution or ionic or nonionic contrast medium. The latter three have known anticoagulant properties. Catheters were flushed with fresh blood and their contents collected at 1-, 2- and 3-min intervals. Blood clots were detected in 65% of catheters flushed with physiologic saline solution at 1 min, 3% of catheters flushed with heparinized saline solution (1, IU/l,OOO ml) at 2 min and 46% at 3 min. This in vitro study verifies that blood clots form rapidly even when anticoagulant agents are flushed into sheaths and much faster if no anticoagulant agent is used. Determinants of thrombus formation. The major determinant of thrombus formation in this study was failure to flush the sheath between each biopsy attempt. The second major determinant was the number of biopsy attempts. Treatment with aspirin or dipyridamole was not protective. No patients in the study were taking warfarin. The experience of the operator did not correlate with the incidence of clot formation. Clinical effects. There did not appear to be any harmful sequelae from these thrombi. They were generally small and it is possible that their composition differs from that of spontaneously occurring venous thrombi that subsequently embolize to the pulmonary circulation (7-12). Nevertheless, they retained their structural integrity during passage through the right side of the heart, a process that occasionally required 1 to 2 min and in one case required hours. Therefore, it is doubtful that they lysed immediately on lodging in the pulmonary vascular bed. Thrombi were frequently observed during aspiration of the sheaths between biopsy attempts and these thrombi had evidence of structural integrity when examined grossly. It is thus likely that the right heart thromboemboli identified echocardiographically did obstruct blood flow to a portion of the pulmonary vascular bed but caused no symptoms because of their small size. Pulmonary perfusion scans and arterial blood gas determinations were not obtained because all patients remained asymptomatic. In heart transplant recipients, the risk of embolization during endomyocardial biopsy appears to be small because a patent foramen ovale is routinely closed at the time of transplantation. Whether these emboli could contribute to the slight increase in pulmonary pressure and resistance noted in transplant recipients is unclear (13,14). However, patients in the general population have a 25% to 35% chance of having a patent foramen ovale (15). Thus, there is a theoretic risk of a paradoxic embolus in patients with cardiomyopathy undergoing endomyocardial biopsy. Conclusions. Small thrombi form rapidly in intravenous sheaths during endomyocardial biopsy and may be dislodged when the bioptome is advanced through the sheath; they are readily visualized by two-dimensional echocardiography. This study confirms the widely held belief that small pulmonary thromboemboli are frequently not associated with symptoms. Furthermore, meticulous flushing of sheaths is necessary during endomyocardial biopsy in patients with cardiomyopathy to avoid paradoxic embolization. We thank Thomas Rector, PhD for assistance with statistical analysis; Dana Souther. BS for typing the manuscript; the nursing staff and cardiology fellows at the University of Minnesota for technical assistance; and the transplant patients at the University of Minnesota for generously agreeing to be subjects in this study.

5 lacc Vol. 19, No.3 March 1, 1992:581-5 KREHER ET AL. 585 References I. Borow M, Crowley JG. Prevention of thrombosis of central venous catheters. J Thorac Cardiovasc Surg 1986;27: Miller LW, Labovitz AJ, McBride LA, Pennington DG, Kanter K. Echocardiography-guided endomyocardial biopsy. Circulation 1988; 78(suppl I1Il:I1I Strachovsky G, Zeldis SM, Katz S, McNulty-Mackey M. Twodimensional echocardiographic monitoring during percutaneous endomyocardial biopsy. J Am Coil Cardiol 1985 ;6: Parrillo JE. Transvenous endomyocardial biopsy: clinical indications, potential complications, and future applications. Chest 1986;9: Fowles RE, Mason JW. Endomyocardial biopsy. Ann Intern Med 1982; 97: Raininko R, Riihela M. Blood clot formation in angiographic catheters. Acta RadioI199:31: Bolognesi R, Tsialtas D, Manca C. Serial cross-sectional echocardiographic detection of multiform floating right-heart thromboemboli. Eur Heart J 1989;1: Saner HE, Asinger RW, Daniel JA, Elsperger KJ. Two-dimensional echocardiographic detection of right-sided cardiac intracavitary thromboembolus with pulmonary embolism. J Am Coil Cardiol 1984;4: Redish GA, Anderson AL. Echocardiographic diagnosis of right atrial thromboembolism. J Am Coil CardioI1983;1:1l Feiner JM, Churchwell AL, Murphy DA. Right atrial thromboemboli: clinical, echocardiographic and pathophysiologic manifestations. J Am Coli CardioI1984;4: II. Cameron J, Pohlner PG, Stafford EG, O'Brien MF, Bett N, Murphy AL. Right heart thrombus: recognition, diagnosis and management. J Am Coli CardioI1985;5: Maciel BC, Neto JAM, Junior LG, et al. Multiple pulmonary thrombolism diagnosed by two-dimensional echocardiography. Chest 1987;92: Greenberg ML, Uretsky BF, Reddy S, et al. Long-term hemodynamic follow-up of cardiac transplant patients treated with cyclosporin and prednisone. Circulation 1985 ;71 : Corcos T, Tamburino C, Leger P, et al. Early and late hemodynamic evaluation after cardiac transplantation: a study of 28 cases. J Am Coil CardioI1988;11: Hagen PT, Scholz DG, Edwards WD. Incidence and size of patent foramen ovale during the first 1 decades of life: an autopsy study of 965 normal hearts. Mayo Clin Proc 1984;59:17-2.