Valve Replacement in Patients With Endocarditis and Acute Neurologic Deficit

Transcription

1 Valve Replacement in Patients With Endocarditis and Acute Neurologic Deficit A. Marc Gillinov, MD, Rinoo V. Shah, MD, William E. Curtis, MD, R. Scott Stuart, MD, Duke E. Cameron, MD, William A. Baumgartner, MD, and Peter S. Greene, MD Division of Cardiac Surgery, The Johns Hopkins Medical Institutions, Baltimore, Maryland Background. Acute neurologic deficits occur in up to 40% of patients with left heart endocarditis. Appropriate evaluation and management of patients with acute neurologic dysfunction who require valve operations for endocarditis remain controversial. This retrospective review was undertaken to develop recommendations for the evaluation and treatment of these challenging patients. Methods. From 1983 to 1995, 247 patients underwent operations for left heart native valve endocarditis at the Johns Hopkins Hospital. From a review of medical and pathology records, 34 patients (14%) with preoperative neurologic deficits were identified. Data on these 34 patients were recorded and analyzed. Results. Causes of neurologic dysfunction included embolic cerebrovascular accident (n = 23, 68%), embolic cerebrovascular accident with hemorrhage (n = 4, 12%), ruptured mycotic aneurysm (n = 3, 9%), transient ischemic attack (n = 2, 6%), and meningitis (n = 2, 6%). Preoperative diagnostic studies included computed tomography (32 patients), magnetic resonance imaging (11 patients), cerebral angiogram (14 patients), and lumbar puncture (2 patients). Computed tomography demonstrated structural lesions in 29 of 32 patients; in only 1 patient did magnetic resonance imaging reveal a lesion not already seen on computed tomography. Of 14 patients having cerebral angiograms, 7 had a mycotic aneurysm. Three mycotic aneurysms had ruptured, and these were clipped before cardiac operations. The mean interval from onset of neurologic deficit to cardiac operation was days for all patients and days for those with embolic cerebrovascular accident. The hospital mortality rate was 6%. New or worse neurologic deficits occurred in 2 patients (6%). Conclusions. Neurologic deficits are common in patients with endocarditis referred for cardiac operations. Despite substantial preoperative morbidity, most of these patients do well if the operation can be delayed for 2 to 3 weeks. Computed tomography scan is the preoperative imaging technique of choice, as routine magnetic resonance imaging and cerebral angiogram are unrewarding. Cerebral angiogram is indicated only if computed tomography reveals hemorrhage. (Ann Thorac Surg 1996;61: ) esults of valve operations in patients with active R infective endocarditis are good, but a difficult management issue that persists is the patient with endocarditis and an acute neurologic deficit [1, 2]. Neurologic dysfunction complicates the course of 10% to 40% of patients with left heart endocarditis [3-5]. The spectrum of neurologic events includes embolic cerebrovascular accident (CVA) (also referred to as ischemic stroke or cerebral septic infarction), embolic CVA with hemorrhage, ruptured mycotic aneurysm, transient ischemic attack (TIA), meningitis, encephalopathy, brain abscess, headache, peripheral neuropathy, and seizure [4-6]. Of these, embolic CVA is the most common [4-6]. Preoperative evaluation and management of the patient with a neurologic deficit and infective endocarditis are controversial. Most authorities recommend computed tomography (CT) scan as the initial diagnostic procedure of choice, but there is disagreement over the utility of magnetic resonance imaging (MRI) and cerebral Presented at the Forty-second Annual Meeting of the Southern Thoracic Surgical Association, San Antonio, TX, Nov 9-11, Address reprint requests to Dr Greene, The Johns Hopkins Hospital, Blalock 618, Baltimore, MD angiography [5]. Although the cerebral angiogram is the only test that can reliably identify mycotic aneurysm, the importance of identifying an unruptured mycotic aneurysm before cardiac operation has been questioned because the indications for operative treatment of these aneurysms are unclear [7-9]. The safety of cardiopulmonary bypass (CPB) in the patient with active endocarditis and an acute neurologic deficit has been debated in the literature [10-14]. Cardiopulmonary bypass may exacerbate neurologic deficits by several mechanisms [14]. Anticoagulation therapy with heparin may extend hemorrhagic infarction and convert nonhemorrhagic CVA to hemorrhagic CVA. Hypotension during CPB may aggravate preexisting ischemic neurologic injury. Finally, CPB may potentiate cerebral edema in areas of blood-brain barrier disruption [14]. Previous studies have examined the operative outcome of patients with endocarditis and with hemorrhagic and nonhemorrhagic CVA, but a clear management plan for such patients is lacking [11]. Questions concerning timing of the operation and the preoperative evaluation of patients with intracerebral hemorrhage and mycotic an by The Society of Thoracic Surgeons /96/$15.00 Published by Elsevier Science Inc PII S (96)

2 1126 G1LLINOV ET AL Ann Thorac Surg STROKE IN ACUTE ENDOCARDIT1S 1996;61: eurysm remain unanswered. The objective of this retrospective study was to develop recommendations for the diagnosis and treatment of patients with left heart native valve endocarditis and preoperative neurologic deficit. Material and Methods We reviewed the hospital records of all patients undergoing valve operations at the Johns Hopkins Hospital from 1983 to During this period, 247 patients had cardiac operations for endocarditis. Each patient satisfied the strict case definition for endocarditis set forth by Von Reyn and colleagues [15], and valve histopathology studies confirmed endocarditis in all cases. Of these 247 patients, 34 had had one or more of the following five preoperative neurologic events: embolic CVA, embolic CVA with hemorrhage, ruptured mycotic aneurysm, TIA, or meningitis. Patients with diffuse encephalopathy, headache, seizure, or peripheral neuropathy but without one of the five conditions listed above were not included in this study. Data collected on the 34 study patients included age, sex, valve(s) involved, organism(s), probable cause of endocarditis, operative indication, type of neurologic deficit, diagnostic studies, interval between neurologic event and cardiac operation, type of valve prosthesis placed, operative outcome, postoperative neurologic status, and postoperative cerebral imaging studies. Unless otherwise indicated, all data are expressed as mean +_ standard error of the mean. Results In the study group of 34 patients, the mean age was 38.3 _+ 2.2 years (range, 13 to 73 years). There were 21 men and 13 women. The causes of endocarditis included intravenous drug abuse (16 patients, 47%), dental procedure (7 patients, 20%), hemodialysis (2 patients, 6%), acupuncture (1 patient, 3%), and unknown (8 patients, 24%). Affected valves were the aortic valve (18 patients), mitral valve (13 patients), and both the aortic and mitral valves (3 patients). Causative microorganisms were predominantly staphylococcal and streptococcal species: Staphylococcus aureus, 13 (38%), Staphylococcus epidermidis, 3 (9 ), Streptococcus viridans, 12 (35%), Streptococcus faecalis, 2 (6%), and others, 4 (12%). Causes of neurologic deficits were as follows. Ruptured mycotic aneurysm occurred in 3 (9%), TIA in 2 (6%), and meningitis in 2 (6%). Overall, 27 patients (80%) suffered an embolic CVA, 4 patients (12%) with hemorrhage and 23 patients (68%) without hemorrhage. These diagnoses were established by CT scan (32 patients), MRI (11 patients), cerebral angiogram (14 patients), and lumbar puncture (2 patients). Results of CT scans were available in 32 patients; I patient with meningitis did not have a CT scan, and the scan from an outside hospital was unavailable for 1 patient with amaurosis fugax. Computed tomography demonstrated structural lesions in 29 of 32 patients scanned. All patients with embolic CVA and 3 of 4 patients suffering embolic CVA with hemorrhage had lesions on CT. Negative CT scans were seen in 1 patient with a TIA, 1 patient with meningitis, and I patient who 7 days later had a small embolic CVA with hemorrhage documented by MRI. Results of MRI in 11 patients confirmed the CT findings in all except the one case mentioned, in which early CT scan failed to show an embolic CVA with hemorrhage. Following the recommendations of consulting neurologists, we obtained cerebral angiograms in 14 patients; these included 6 of 7 patients with evidence of hemorrhage on CT and 8 patients with no hemorrhage on CT. Of the 6 patients with hemorrhage, 3 were found to have ruptured mycotic aneurysms. One patient with an embolic CVA and hemorrhage was thought to have an incidental unruptured Berry aneurysm. This aneurysm was resected before the patient had cardiac operation, and pathologic study showed that it was an unruptured mycotic aneurysm. Of the 8 patients without intracerebral hemorrhage who had a cerebral angiogram, 3 had unruptured mycotic aneurysms; none of these aneurysms was treated operatively. Once the diagnosis of endocarditis was established, all patients received intravenous antibiotics. Indications for operation were hemodynamic deterioration (24 patients), persistent sepsis (6 patients), and recurrent systemic emboli (4 patients). Although most patients could wait at least 2 weeks before operation, 7 patients had severe hemodynamic instability and required urgent valve replacement within 7 days of their neurologic event. The mean interval from onset of neurologic dysfunction to cardiac operation was days (range, I to 55 days), and the median interval was 17 days. This interval was days for patients with embolic CVA, days for patients suffering embolic CVA with hemorrhage, days for patients with ruptured mycotic aneurysm, days for patients with TIA, and days for patients with meningitis. The 3 patients with ruptured mycotic aneurysms confirmed by angiogram underwent clipping of the aneurysm before cardiac operation. In 2 patients, hemodynamic performance was such that valve replacement could be delayed until there was no evidence of cerebral edema by CT scan (21 and 34 days); both of these patients survived valve replacement and had resolution of their neurologic deficits. One patient had successful clipping of his mycotic aneurysm but required urgent cardiac operation 2 days later because of hemodynamic deterioration. Although CT performed after valve replacement revealed no new lesion, he suffered from seizures and encephalopathy and died of sepsis and multisystem organ failure 1 month later. The choice of valve prosthesis depended upon patient factors and surgeon preference. In general, patients judged to be medically compliant and without a history of bleeding problems received mechanical valves. Seventeen patients received mechanical valves and 16 patients had bioprostheses. In 1 patient, aortic root replacement with a homograft was performed for extensive destruction of the aortic root. There were two operative deaths, for a hospital mortality rate of 6%. The first death occurred in the patient

3 Ann Thorac Surg GILLINOV ET AL ;61: STROKE IN ACUTE ENDOCARDITIS mentioned earlier who required urgent valve operation 2 days after his mycotic aneurysm was clipped. The second death occurred in a 30-year-old man with multiple nonhemorrhagic left cerebral infarctions documented by preoperative CT scan. Six weeks after his CVA, he underwent uneventful aortic valve replacement. He did well postoperatively and had no new neurologic deficit. He was discharged on postoperative day 13. Four days after discharge, he was found dead at home. Autopsy did not reveal the cause of death. The other 22 patients with embolic CVA survived their operations and were discharged from the hospital. All 4 patients with embolic CVA and hemorrhage tolerated valve replacement. New or worse neurologic deficits occurred in 2 patients after valve operation. As noted previously, the patient who underwent cardiac operation 2 days after clipping of a mycotic aneurysm experienced new seizures and encephalopathy of uncertain cause. The second patient with worsened neurologic status was a 64-year-old woman who underwent mitral valve replacement 12 days after a nonhemorrhagic embolic CVA in the right frontal lobe. Postoperatively, she had new symptoms consistent with posterior circulation stroke. Neither CT nor MRI revealed a new lesion, but her neurologists were confident that her findings were consistent with a new embolic CVA. Of the other 31 operative survivors, 29 had neurologic improvement by the time of discharge and 2 had no change in neurologic status. Comment Neurologic complications occur in 10% to 40% of patients with endocarditis and are associated with increased mortality rates [3-6]. Neurologic manifestations of infective endocarditis include embolic CVA, embolic CVA with hemorrhage, TIA, ruptured mycotic aneurysm, meningitis, seizure, headache, nonfocal encephalopathy, and peripheral neuropathy [3-6]. Of these, embolic CVA without hemorrhage is the most common [4-6]. Neurologic complications of endocarditis tend to occur early in the course of infection and are more common with Staphylococcus aureus endocarditis and uncontrolled sepsis [3]. Recurrent emboli are rare once infection is controlled [3]. Although infective endocarditis is often cured with antibiotics, many patients still require operation during the active phase of infection [1, 2]. The patient with an acute neurologic deficit who requires valve operation for endocarditis presents a special management challenge. Anticoagulation therapy, hypotension, and edema associated with CPB may exacerbate neurologic injury. Thus, guidelines for patient evaluation and timing of operation are necessary. When valve replacement was performed 5 days after acute nonhemorrhagic cardiogenic embolism, Maruyami and associates [12] noted severe neurologic deterioration in 4 of 14 patients (29%). They suggested that infective endocarditis, a large cerebral infarct, and a short interval between cerebral injury and cardiac operation predispose to hemorrhagic transformation and worse neuro- logic injury. Similarly, Matsushita and co-workers [13] reported that cardiac operation for endocarditis performed within 5 days of a cerebral embolic event was frequently associated with neurologic deterioration. They concluded that an interval of at least 11 days between the CVA and valve replacement is advisable. They also recommended that this interval be extended to 4 weeks for embolic CVA with hemorrhage. In a series of patients suffering primarily from active endocarditis, Zisbrod and colleagues [14] performed valve replacement on 12 patients at an average of 12.7 days after embolic CVA. One patient died, and there were no new strokes. The largest series examining valve replacement in patients with endocarditis and embolic CVA is that of Ting and associates [10] from the University of Illinois. They reviewed the outcomes in 45 patients with recent cerebral emboli who required valve replacement. In this series, the mean interval from onset of neurologic deficit to cardiac operation was 11.4 days. No patient with nonhemorrhagic embolic CVA sustained a perioperative stroke. Only preoperative hemorrhage on CT scan correlated with mortality and perioperative stroke. Surprisingly, the timing of valve replacement was not related to outcome. When considering these data, it is important to realize that a large percentage of these patients were asymptomatic and had embolic CVA detected by screening CT scan. Therefore, the true interval from cerebral embolism to cardiac operation was unknown in these patients. Furthermore, the interval between onset of neurologic deficit and valve replacement for symptomatic patients was not reported, so specific recommendations for timing of operation could not be made. As in other series, we relied primarily on CT scan for the diagnosis of neurologic complications of endocarditis [10, 12-14]. Thirty-two of 34 patients had CT scans. Computed tomography identified embolic CVA in all 23 patients with that diagnosis, embolic CVA with hemorrhage in 3 of 4 patients, and cerebral hemorrhage in all 3 patients with ruptured mycotic aneurysm. As noted previously, the single false-negative CT scan may have been performed too soon after cerebral infarction to detect any changes. These results confirm those of Ting and associates [10], suggesting that CT should be the first diagnostic test for evaluating the patient with infective endocarditis and neurologic deficit. Although 11 patients also had MRI, in most instances MRI served only to confirm the results of CT. Thus, MRI has a limited role in the workup of these patients. The place for cerebral angiography in the evaluation of patients with infective endocarditis is controversial. Cerebral angiography is the only test that can reliably document a mycotic aneurysm, with a sensitivity of approximately 90% [7-9]. It is estimated that cerebral mycotic aneurysm develops in 1% to 12% of patients with infective endocarditis and that 10% of these aneurysms rupture [4, 5, 7-9]. In large series, 0.5% to 2% of patients with infective endocarditis suffered a ruptured cerebral mycotic aneurysm [4-9]. Serial angiographic studies demonstrated that at least 50% of mycotic aneurysms resolve with appropriate antibiotic therapy [9, 16]. With

4 1128 GILLINOV ET AL Ann Thorac Surg STROKE IN ACUTE ENDOCARDITIS 1996;61: available diagnostic techniques, however, it is currently impossible to determine which mycotic aneurysms will rupture and which will resolve [9, 16]. Furthermore, data from the Cleveland Clinic demonstrated that the risk of late rupture of an unsuspected mycotic aneurysm after a complete course of antibiotics is low [8]. Using data from published series and mathematic modeling, Van der Meulen and co-workers [9] found no benefit of routine cerebral angiography in patients with infective endocarditis and neurologic deficit. Thus, routine cerebral angiography for all patients with endocarditis and neurologic deficit does not seem warranted [5, 9]. Although routine cerebral angiography may not be indicated, DiSesa [11] has suggested that cerebral angiograms be obtained in patients with intracranial hemorrhage. In our series, cerebral angiograms were obtained in 14 patients. Of 6 patients with intracerebral hemorrhage, 3 were found to have ruptured mycotic aneurysms, and these were clipped before cardiac operation. Three of 8 patients without cerebral hemorrhage had unruptured mycotic aneurysms. None of these aneurysms was treated operatively, and all 3 patients had uneventful valve replacement. Given these results, we recommend cerebral angiography only if CT reveals intracerebral hemorrhage. The mean interval from onset of neurologic dysfunction to valve replacement was 22.2 _+ 2.8 days. This interval was _ 3.0 days for the largest group of patients, those with embolic CVA. As noted previously, other groups have demonstrated that most patients do well if the interval between cerebral embolism and valve replacement is approximately 2 weeks [10, 13, 14]. Unfortunately, there are no prospective, randomized studies examining strategies for timing of operation in these patients. Thus, decisions must be based on retrospective series and the surgeon's assessment of individual patients. Certain general recommendations are nevertheless possible. Because mitral valve vegetations greater than 10 mm in diameter increase the risk of embolic events, patients with this finding might be considered for early cardiac operation [17]. In addition, patients who do not respond to medical therapy and experience intractable heart failure, persistent sepsis, or paravalvular abscess will usually require urgent operation. When possible, however, we recommend that cardiac operation be delayed 2 to 3 weeks for patients who have nonhemorrhagic embolic CVA. Few data exist to develop recommendations for the management of patients suffering embolic CVA with hemorrhage. It is clear, however, that these patients are at increased risk for perioperative stroke. Ting and associates [10] operated upon 5 patients with cerebral hemorrhage, and all had perioperative stroke. The interval between cerebral hemorrhage and valve replacement in these patients was not stated. Matsushita and colleagues [13] performed valve replacement in 6 patients with intracerebral hemorrhage a minimum of 23 days from the time of hemorrhage. There were no perioperative strokes or deaths. In our series, 4 patients with embolic CVA and hemorrhage had cardiac operations an average of days from the time of cerebral hemorrhage; there were no deaths or new neurologic deficits. Therefore, in stable patients with embolic stroke and hemorrhage, we recommend an interval of 4 weeks between the neurologic event and cardiac operation. As noted earlier, the indications for cerebral angiography and the treatment of mycotic aneurysm are controversial. Most authorities agree that a ruptured mycotic aneurysm should be resected or clipped before cardiac operation [18, 19]. This was our strategy. Two of 3 patients did well with staged operative treatment of ruptured mycotic aneurysm followed by valve replacement; the third patient died after emergent cardiac operation was performed only 2 days after the mycotic aneurysm was clipped. In the 2 survivors, valve replacement was performed after neurologic status had stabilized and cerebral edema had subsided (21 and 34 days). Others have reported successful outcomes when cardiac operation was performed 8 and 14 days after aneurysm clipping [18, 19]. We recommend an interval of 2 to 3 weeks between aneurysm clipping and valve replacement in stable patients. Cardiac operations should be performed only when neurologic status has stabilized and CT scan reveals resolution of cerebral edema and no ongoing cerebral hemorrhage. The approach to the patient with unruptured mycotic aneurysms is also controversial. In general we obtain neurosurgical consultation on all such patients. It is not possible to predict the natural history of these lesions, and there is no evidence that CPB or heparin treatment predisposes these lesions to rupture [5, 16]. Our current policy is not to clip or resect unruptured mycotic aneurysms before undertaking valve replacement. Further recommendations must await studies of the natural history of cerebral mycotic aneurysms. We have developed an algorithm for the evaluation and management of the endocarditis patient with a neurologic deficit (Fig 1). Although prospective studies of such patients do not exist and the numbers of patients analyzed in this and other reports are small, certain general guidelines can be drawn. Computed tomography is the diagnostic procedure of choice. Cerebral anglogram is used if CT reveals intracerebral hemorrhage. Lumbar puncture is used to establish the diagnosis of meningitis. If possible, valve replacement is delayed 2 to 3 weeks in patients with embolic CVA and 4 weeks in patients with embolic CVA and hemorrhage. Ruptured mycotic aneurysms are clipped or resected before cardiac operation. There is no evidence that patients with TIA or meningitis benefit from a delay in operation; they may undergo valve replacement once the CT scan confirms the absence of cerebral infarct or hemorrhage. Neurologic complications are common in patients with infective endocarditis. Herein, we present guidelines for the diagnosis and management of several of the most common neurologic manifestations of endocarditis. These recommendations should minimize unnecessary diagnostic tests and result in good outcomes for most of these challenging patients.

5 Ann Thorac Surg GILLINOV ET AL ;61: STROKE IN ACUTE ENDOCARDITIS Embolic ~ Monitor ~ Valve CVA v 2-3 Weeks v Surgery Embolic CVA ~ Monitor ~ Valve with Hemorrhage 4 Weeks Surgery Neurologic Deficit l and ~ CTScan a~..~"! ~ Intracranial ' ~ ~ beredral -, / Active Endocarditis ~ v Hemorrhage Angiogram N rosur er Ruptured eu g Y: Monitor Valve I I Mycotic ~Clip/Resect ~ ~ Aneurysm Aneurysm 2-3 Weeks Surgery Negative %.~ Suspect TIA ~ Valve Surgery, 7 without uejay Suspect ~ Lumbar ~Positive ~Valve Surgery Meningitis Puncture Without Delay Fig 1. Algorithm for the diagnosis and management of patients with acute neurologic dejicit requiring cardiac operations for infective endocarditis. (CT = computed tomography; CVA = cerebrovascular acciden# TIA = transient ischemic attack.) We thank Barbara Fleischman for her assistance in preparation of the manuscript. References 1. David TE, Bos J, Christakis GT, Brofman PR, Wong D, Feindel CM. Heart valve operations in patients with active infective endocarditis. Ann Thorac Surg 1990;49: Miller DC. Determinants of outcome in surgically treated patients with native valve endocarditis (NVE). J Cardiac Surg 1989;4: Hart RG, Foster JW, Luther MF, Kanter MC. Stroke in infective endocarditis. Stroke 1990;21: Salgado AV, Furlan AJ, Keys TF, Nichols TR, Beck GJ. Neurologic complications of endocarditis: a 12-year experience. Neurology 1989;39: Jones HR, Siekert RG. Neurologic manifestations of infective endocarditis: review of clinical and therapeutic challenges. Brain 1989;112: Kanter MC, Hart RG. Neurologic complications of infective endocarditis. Neurology 1991;41: Morawetz RB, Karp RB. Evolution and resolution of intracranial bacterial (mycotic) aneurysms. Neurosurgery 1984; 15: Salgado AV, Furlan AJ, Keys TF. Mycotic aneurysm, subarachnoid hemorrhage, and indications for cerebral angiography in infective endocarditis. Stroke 1987;18: Van der Meulen JHP, Weststrate W, Van Gijn J, Habbema JDF. Is cerebral angiography indicated in infective endocarditis. Stroke 1992;23: Ting W, Silverman N, Levitsky S. Valve replacement in patients with endocarditis and cerebral septic emboli. Ann Thorac Surg 1991;51: DiSesa VJ. Art and science in the management of endocarditis. Ann Thorac Surg 1991;51: Maruyama M, Kuriyama Y, Sawada T, Yamaguchi T, Fujita T, Omae T. Brain damage after open heart surgery in patients with acute cardioembolic stroke. Stroke 1989;20: Matsushita K, Kuriyama Y, Sawada T, et al. Hemorrhagic and ischemic cerebrovascular complications of active infective endocarditis of native valve. Eur Neurol 1993;33: Zisbrod Z, Rose DM, Jacobowitz IJ, Kramer M, Acinapura AJ, Cunningham JN. Results of open heart surgery in patients with recent cardiogenic embolic stroke and central nervous system dysfunction. Circulation 1987; 76(Suppl 5): Von Reyn CF, Levy BS, Arbeit RD, Friedland G, Crumpacker CS. Infective endocarditis: an analysis based on strict case definitions. Ann Intern Med 1981;94: Brust JCM, Dickinson PCT, Hughes JEO, Holtzman RNN. The diagnosis and treatment of cerebral mycotic aneurysms. Ann Neurol 1990;27: Mugge A, Daniel WG, Frank G, Lichtlen PR. Echocardiography in infective endocarditis: reassessment of prognostic implications of vegetation size determined by the transthoracic and transesophageal approach. J Am Coil Cardiol 1989;14: Jara FM, Lewis JF, Magilligan DJ. Operative experience with infective endocarditis and intracerebral mycotic aneurysm. J Thorac Cardiovasc Surg 1980;80: Kuki S, Yoshida K, Suzuki K, Matsumura R, Okuda A. Successful surgical management for multiple cerebral mycotic aneurysms involving both carotid and vertebrobasilar systems in active infective endocarditis. Eur J Cardiothorac Surg 1994;8:

6 1130 GILL1NOV ET AL Ann Thorac Surg STROKE IN ACUTE ENDOCARDITIS 1996;61: DISCUSSION DR NORMAN J. SNOW (Cleveland, OH): You have not mentioned the use of transesophageal echocardiography in this disease. In this day, when echocardiographers visualize the pedunculated mass or vegetation on the valve, they are encouraging us to operate right away. And in a very small number of patients, fewer than 6, we have operated within 24 to 48 hours of a stroke, if the patients do not have a hemorrhagic stroke, with basically the same results and with no worsening of neurologic deficit. How do you persuade your medical colleagues to allow you to wait for 22 days to do this operation? DR GILLINOV: This was a retrospective study, and 22 days was the mean interval between the neurologic event and operation. Seven patients had their operation within 1 week of the neurologic event. The importance of vegetations seen on echocardiography remains controversial in the echocardiography and cardiology literature. It is virtually impossible to predict which vegetations will embolize and lead to stroke. There are data that vegetations on the mitral valve greater than 10 mm in size significantly increase the risk of suffering an embolic event, and in these patients we would be inclined to operate more urgently. But in a patient who is stable and who does not have mobile or pedunculated lesions by echocardiography, we do try to wait at least 10 to 15 days before operating. DR SAFUH ATTAR (Baltimore, MD): What type of valve was used? Did you use anticoagulation postoperatively? In addition, were there any cases of recurrent endocarditis and, if so, did they require reoperation? What was the result? DR GILLINOV: In this series, 17 patients received mechanical valves, 16 had bioprosthetic valves, and 1 had an aortic root homograft placed for extensive destruction of the root. The choice of valves was dictated in large part by the fact that nearly half of these patients were intravenous drug abusers and would not be candidates to receive a mechanical valve. In most eases, people who would be compliant, who did not have a bleeding diathesis, and who had not had a cerebral hemorrhage would receive a St. Jude valve, because of data from the Texas Heart Institute suggesting a reduced incidence of recurrent endocarditis. As far as follow-up goes, unfortunately, because so many of these patients were intravenous drug abusers, they were hard to locate for follow-up. We do know of 2 cases of recurrent endocarditis after the study; I patient underwent re-replacement of the valve and did well, and the other died.