Septic embolism complicating infective endocarditis David Kitts, M D , Frederic S. Bongard, M D , and Stanley R. Klein, M D ,
Torrance, Calif. Embolic phenomena in patients with infective endocarditis may complicate the placement of a cardiac vah~lar prosthesis. To evaluate the vascular consequences of these emboli, a 15-year review of 102 patients undergoing valve replacement for proven infective endocarditis was undertaken. Thirty-one patients with 36 episodes of septic embolization were identified. Ten of these were separate extremity occlusive events. All patients with extremity emboli were admitted with pain; four had limb-threatening emboli. All patients grew gram-positive bacteria from their blood except a single Candida albicans isolate. Appropriate antimicrobial therapy was used in all patients. Angiography confirmed the diagnosis in 11 of 12 patients. Embolic targets included the lower extremities in all except a single instance. Four patients had multiple emboli. All but one of the vascular procedures were carried out subsequent to or simultaneously with cardiac valve replacement. Initial operative management included embolectomy (4) and primary amputation (2). Two delayed procedures were required. One patient died. Four patients had limited ischemia that resolved with antibiotics and anticoagulation. This report suggests that infective endocarditis requiring valvaxlar replacement is associated with embolization in one third of patients. The presentation of peripheral vascular emboli is that of acute extremity ischemia. The diagnosis should be confirmed by angiography to rule out the possibility of multiple emboli. When possible, valve replacement should precede peripheral vascular management, which may include operative or medical components as dictated by the extent of limb ischemia. (J VASe SuRe 1991;14:480-7.)
Vascular surgeons are increasingly being called on to recognize the morbidity imposed by coexistent cardiac disease in patients seeking treatment for extremity ischemia. In this regard, few patients will present as challenging a problem as the patient with infective endocarditis (IE) and extremity ischemia caused by septic embolism. Management may be complicated by poor cardiac status as well as the effects of additional emboli to the cerebral, visceral, and other peripheral vascular beds. 1-4 Patients with IE referred for cardiac valve replacement (CVR) are a source of such management problems because many have experienced multiple embolic events as well as congestive heart failure. The purpose of this report is to delineate the key factors leading to From the Departmentof Surgery,Harbor-UCLAMedicalCenter, Torrance. Presented at the Sixth AnnualMeeting of the Western Vascular Society, Palm Springs, Calif.,Jan. 13-16, 1991. Reprint requests: StanleyR. Klein,MD, Department of Surgery, Harbor-UCLA MedicalCenter, 1000 W. Carson St., Box 15, Torrance CA 90509. 24/6/29914 480
successful outcome and minimal morbidity in the management of these patients with IE and peripheral septic embolism. METHODS
Cardiac valve replacement for proven IE was performed in 102 patients at Harbor-UCLA Medical Center between January 1974 and December 1989. Thirty-one of these patients (30%) had septic embolic events: nine patients had cerebral emboli, six had visceral emboli, six had isolated pulmonary emboli, and 10 patients had major peripheral vascular embolization. The characteristics and management of the latter 10 patients with primarily peripheral vascular emboli form the basis of this report. PATIENTS The 10 patients in this report include six men and four women with ages ranging from 20 to 44 years at the time of initial presentation. Individual predisposing conditions, clinical course, and management are outlined in Table I. A history of recent intravenous drug abuse was the predominant risk factor in
Volume 14 Number 4 October 1991
Septic embolism complicating infective endocarditis
481
Table I. Individual predisposing conditions, clinical course, patient management Patient (age)
Risk factor
1. L.G. (31)
IVDA, bicuspid AV
2. M.D. (24)
IVDA
3. C.T. (24)
prior MVR
4. M.M. (25)
IVDA
5. D.R. (32)
IVDA
6. A.L. (44)
IVDA
7. R.M. (43)
IVDA
8. R.J. (23)
Septic TAB
9. R.B. (26)
Dental work
10. A.B. (20)
IVDA
Symptoms on admission
Valve
Embolism
Culture
Management
Fever, CHF, bilateral ischemic leg pain Fever, CHF, abdominal pain, ischemic right leg pain Fever, ischemic left foot pain
Aortic
Bilateral external lilac arteries
Aortic
Right SFA
Mitral
Left femoral/popliteal embotectomy
Good
Fever, LUQ abdominal pain, ischemic left foot pain Fever, CHF, hypotension, ischemic left foot pain Fever, CHF, hypotension Fever, CHF, left facial paralysis, ischemic left leg pain Fever, hypotension, LU, RL extremity paresthesias Fever, ischemic left leg pain Fever, CHF, weight loss
Aortic/mitral
Left common s. epi femoral artery, left popliteal artery, right renal artery Left popliteal S. viridans artery, spleen, right MCA
AVR/MVR, splenectomy, delayed femoral distal bypass
Good
Aortic mitral
Left tibial artery
S. faecalis
AVR/MVR, delayed transmetatarsal amputation
Good
Aortic
Spleen left popliteal artery Left common iliac artery, right MCA aneurysm
AVR, splenecS. viridans tomy, delayed C. albicans left BKA AVR, Coumadin S. viridans
Died
Aortic
Candida AVR parapsilosis Bilateral femoral/embolectomy Streptococcus AVR, right SFA viridans embolectomy
Outcome
Good
Good
Good
Mitral
Left brachial Group B artery, right streptopopliteal cocci artery, spleen
MVR, Coumadin
Good
Aortic
Left tibial artery
S. viridans
AVR, Coumadin, antibiotics
Good
Mitral
Right cornmon lilac artery, left superficial femoral artery
S. viridans
MVR, Coumadin
Good
IVDA, Intravenous drug abuse; CHF, congestive heart failure; LV, left upper; AVR, aortic valve replacement; LUQ, left upper quadrant; TAB, therapeutic abortion; RL, right lower; SFA, superficial femoral artery; MCA, middle cerebral artery; BKA, below-knee amputation.
this group, occurring in 7 of 10 patients (70%). Exceptions included a 24-year-old woman (case 3) with porcine mitral valve replacement 1 year previously, a 23-year-old woman (case 8) with septic endometritis 5 days after therapeutic abortion, and a 26-year-old man (case 9) with recent dental extractions.
CLINICAL PRESENTATION All patients were admitted with fever, and all but two (80%) had ischemic extremity pain. Embolic occlusive events after CVR developed in both patients without ischemia on admission (cases 6 and 8). Congestive heart failure with pulmonary edema was present in 50% of patients, and three patients
482
Journal of VASCULAR SURGERY
IGtts, Bongard, and IGein
Table II. Site of embolism (10 patients) Site (artery)
No. of events
Common iliac External iliac Common femoral Superficial femoral Popliteal Tibial Brachial
2 2 1 2 4 2 1
Total
14
(30%) had hypotension with systolic blood pressure less than 90 mm Hg on admission. In addition, one patient (case 7), had a dense neurologic deficit as a result of cerebral aneurysm, and another had abdominal pain caused by splenic embolism. Overall, 50% of patients were demonstrated to have remote multiorgan emboli in addition to their primary vascular event. Three patients had concomitant splenic emboli (cases 4, 6, and 8), two patients had cerebral lesions (cases 4 and 7), and a single patient (case 3), had a right renal embolus. BACTERIOLOGY The source of septic embolism was an isolated aortic valvular lesion in five patients, both aortic and mitral involvement in two patients, and isolated mitral valvular infection in the remaining three patients. All patients had at least two positive blood cultures on separate occasions. Gram stain and culture of valvular vegetations demonstrated the offending organism in 100% of cases. Eight of the nine patients with bacterial endocarditis (89%) had streptococcal infections, the only exception being the prosthetic valve infection caused by Staphylococcus epidermidis. A single patient (case 1), was admitted with fungal endocarditis caused by Candida parapsi-
losis. PRESENTATION OF LESIONS
Eight o f the 10 patients in this series were symptomatic from major peripheral vascular emboli at the time of admission. Septic emboli developed in the two additional patients and they became symptomatic shortly after CVR (cases 6 and 10). Initial vascular examination demonstrated a pulse deficit distal to ithe embolic site in 9 of the 10 patients. Neurologic impairment manifested as extremity sensory deficit was present in four patients, and four patients had overt limb threat (cases 1 to 4). All patients underwent angiography before any vascular procedure. Fourteen embolic events were
identified in the 10 patients. Embolic targets were as outlined in Table II. Included were four patients who had multiple simultaneous peripheral emboli. The patients with multiple emboli had lesions of the bilateral external iliac arteries (1), common iliac and contralateral common femoral arteries (1), common femoral and ipsilateral popliteal arteries (1), and popliteal and contralateral brachial arteries (1). All of the angiographically identified lesions were symptomatic. It is of note that there were no asymptomatic lesions detected. MANAGEMENT Treatment of embolic episodes varied considerably between patients and was dependent on cardiac status, magnitude of limb ischemia, and severity of concurrent illness caused by other remote emboli. Patients could be grouped into (1) those requiring emergent CVR and simultaneous embolectomy, (2) patients requiring emergent embolectomy whose cardiac status was sufficiently stable to tolerate a noncardiac procedure, and (3) patients in whom the degree of limb ischemia did not warrant the risk of emergency intervention. EMERGENCY EMBOLECTOMY
Four patients underwent emergency embolectomy. In two patients (cases 1 and 2) this was performed simultaneously with CVR. Both patients were admitted with florid congestive heart failure as a result of aortic regurgitation in addition to limb ischemia. Cardiac status would have prohibited initial embolectomy. Concomitant CVR and embolectomy resulted in excellent outcome in both cases with resolution of ischemia and complete recovery before discharge from the hospital. The third emergency embolectomy (case 3) was performed before CVR for left lower extremity limb threat caused by common femoral artery and popliteal artery embolic occlusion. Intraoperative invasive hemodynamic monitoring of cardiac function allowed for uncomplicated general anesthesia and embolectomy despite mitral insufficiency. The concomitant right renal embolism remained asymptomatic, and the patient had an uneventful complete recovery after subsequent CVR. The final emergency embolectomy was performed in case 4, a 25-year-old man admitted with high fever, left upper quadrant abdominal pain, and an ischemic left lower extremity. Abdominal CT scanning demonstrated a splenic abscess, and angiography revealed isolated left popliteal occlusion. This patient's cardiac status allowed for simultaneous
Volttme 14 Number 4 October 1991
Septic embolism complicating infective endocarditis 483
splenectomy and left popliteal embolectomy resulting in a viable extremity. Subsequent CVR was complicated by left hemiparesis as a result of right middle cerebral embolism. After a prolonged recovery he had left leg claudication necessitating femoral posterior tibial bypass for chronic popliteal occlusion. He was eventually discharged from the hospital with no symptoms and fully recovered 5 months after initial admission. AMPUTATION Two patients were treated with amputation as the initial procedure for peripheral vascular embolism (cases 5 and 6). The first of these, a 32-year-old woman was admitted with congestive heart failure, left foot pain, and absent pedal pulses and required emergent aortic valve replacement/mitral valve replacement for hemodynamic instability. Intraoperatire angiography demonstrated the crural vessels to be patent to the ankle. Despite this, dry gangrene of the left forefoot developed after operation because of distal tibial emboli. She underwent transmetatarsal amputation 6 weeks after CVR and was discharged from the hospital able to walk with a cane. The remaining amputation was performed in a 44-yearold man admitted with congestive heart failure and sepsis who also required emergent aortic valve replacement for progressive deterioration (case 6). His postoperative course was complicated by splenic abscess requiring splenectomy and later multisystem organ failure caused by Candida sepsis. Three weeks after CVR his left foot became pulseless, but ongoing hemodynamic instability precluded intervention. He subsequently underwent a left below-knee amputation as a last resort but died shortly thereafter of progressive multisystem organ failure. NONOPERATIVE MANAGEMENT Nonoperative management consisting of antibiotics and anticoagulation was used in the remaining four patients. In three of the four (cases 7, 8, and 10) the extent of fimb ischemia did not warrant aggressive intervention given the coexistent cardiac insufficiency. One of these patients (case 7) was admitted with congestive heart failure and right middle cerebral stroke in addition to left lower extremity pain. Angiography demonstrated a left common iliac artery occlusion with reconstitution of the common femoral artery. Despite the impressive arteriogram (Fig. 1), his ischemic extremity symptoms completely resolved with antibiotics and anticoagulation before CVR. At follow-up after discharge from the
Fig. 1. Arteriogram shows left common iliac occlusion.
hospital he remained asymptomatic and had regained full femoral and distal pulses. Two additional patients were admitted with ischemic extremity pain without limb threat (cases 8 and 9). In both of these antibiotics and anticoagulation again resulted in complete resolution of symptoms during the period of treatment before CVR. The final patient treated nonoperatively had lower extremity pain 2 days after emergent mitral valve replacement for IE and progressive deterioration. Angiography demonstrated right common iliac artery and left superficial femoral artery occlusion with distal reconstitution. Embolectomy was recommended, but the patient refused operation. He was treated with antibiotics, and sodium warfarin (Coumadin) was continued. Two weeks after this event, he left the hospital against advice and was asymptomatic with exertion. DISCUSSION Excluding congestive heart failure, arterial embolization of infected debris and its sequelae is the most common complication of IE. This event occurs in
484 IGtts, Bongard, and IGein
approximately 30% of patients with acute IE in published series.l-3 We found a similar incidence of septic embolism in our patients with acute IE referred for CVR. Thirty-one of the 102 patients (30%) undergoing CVR for IE had major documented embolic events. The manifestation of such an event depends ultimately on the destination of the embolus. Although pulmonary, splenic, and renal emboli are frequently well tolerated, a single embolus to the cerebral, coronary, or peripheral vascular bed can permanently disable a previously fully functional patient. Unlike the cerebral and coronary circulations, the peripheral vasculature is readily accessible and provides the opportunity to potentially reverse organ ischemia and tissue death. Management may be complicated, however, by coexistent cardiac dysfunction and the added morbidity of simultaneous emboli to the cerebral, visceral, and other peripheral vascular targets. Fifty percent of patients in this series were admitted in congestive heart failure. Half again had multiorgan emboli in addition to limb ischemia, including three splenic, one renal, and two cerebral lesions. This tendency for septic embolization to be recurrent and widespread has been noted previously. Six of eight patients (75%) reported by Nakayama et al.4 and 50% of patients in the series by Pruitt et al.5 suffered multiple systemic emboli. This has serious implications if complications from unrecognized cerebral and visceral emboli are to be avoided. The single death in this series occurred in a patient who underwent emergent aortic valve replacement and subsequently required splenectomy for ongoing sepsis caused by splenic abscess not recognized before operation. Although unavoidable in this instance, a valuable lesson was reiterated. The sensitivity of CT scanning in detecting splenic emboli is well documented, and many authorities believe that this should be routinely performed in patients with IE to avoid such complications.6"7 Multiple peripheral vascular emboli and "embolomycotic" aneurysms may also complicate the course of IE. Seven of nine patients in the series by Dean et al.8 had multiple vascular emboli, and 14 aneurysms were discovered in these nine patients. In the present series, angiography detected multiple vascular emboli in four patients (40%), and embolomycotic aneurysms were detected in two patients (20%). Although all embolic lesions in this series were symptomatic, the propensity for septic emboli and particularly aneurysms to remain silent has been stressed by Nakayama et al.,4 Dean et al.,8 and Freischlag et al.9
Journal of VASCULAR SURGERY
Angiography to date represents the only means of recognizing and treating these potentially silent lesions and has become an integral part of the vascular workup in patients with IE and septic embolism. 4"8"9 The noninvasive vascular laboratory has little experience in this regard. Although embolectomy may be one of the less challenging procedures undertaken by vascular surgeons, treatment of the patient with limb ischemia and IE promises to be a complex problem. Even in the face of limb threat, cardiac performance and hemodynamic stability must remain priorities. Patients admitted while in congestive heart failure caused by valvular incompetence will poorly tolerate any form of anesthesia associated with a lengthy vascular procedure. 1° It is important for the vascular surgeon to know that the endarteritis that rapidly results from the infected embolus frequently destroys the integrity of the arterial wall. Thus what may begin as a "simple embolectomy" often escalates into a more complicated revascularization requiring a much longer period of time. It is also important to recognize that valuable time may be lost during attempted diuresis and afterload reduction to improve cardiac performance. The direct correlation between preoperative delay and subsequent amputation was emphasized by Elliot et al.ll in their review of 225 patients with nonseptic arterial emboli. After 8 hours the amputation rate was 50% higher, and after 24 hours it had doubled. 11 In the setting of cardiac decomposition caused by IE and limb threat as a result of concomitant septic embolism, serious consideration should be given to simukaneous emergent CVR and embolectomy. Cardiac surgeons may be reluctant to proceed with emergent CVR in the presence of acute IE, citing the potential for increased prosthetic valve infection in this setting. The incidence of prosthetic valvular reinfection in acute IE is less than 10%. 12-14 Furthermore, the benefit of limb salvage with simultaneous eradication of the source of additional emboli may outweigh this risk. In those instances in which limb threat is present and CVR is clearly not indicated, invasive intraoperative hemodynamic monitoring and the presence of an experienced anesthetist may prevent cardiac decompensation during a procedure. Finally, the magnitude of limb ischemia may not warrant embolectomy, particularly if the patient's condition is complicated by congestive heart failure or stroke as was the case in one of our patients. Karchner, 15 DiNubile, 16 and others have demonstrated that antibiotics and often anticoagulation
Volume 14 Number 4 October 1991
were frequently sufficient to treat embolic complications o f IE. F o u r patients in this series were treated in this m a n n e r and all had complete resolution o f ischemic symptoms. CONCLUSION Septic embolization occurs in at least 30% o f patients with I E referred for C V R . A high index o f suspicion m u s t be maintained if this complication is to be detected. O n e third o f these embolic events are to the pcripheral vasculaturc. Such emboli arc frequently multiple and m a y be accompanied by mycotic aneurysms, or splenic, cerebral, and renal infarcts. A n g i o g r a p h y and abdominal C T scanning are invaluable in determining before operation the sites o f septic embolism. A low threshold for obtaining these studies is indicated. The noninvasive laboratory m i g h t provide an excellent screening modality in the future. T h e presentation is frequently that o f limb ischemia superimposed o n congestive heart failure. T h e extent o f limb ischemia and degree o f cardiac dysfunction will dictate the priority o f the problems. W h e n possible, C V R should proceed before or simultaneously with e m b o l e c t o m y to avoid the m o r b i d i t y o f e m b o l e c t o m y in a hemodynamically unstable patient.
REFERENCES
1. Wilson WR, Guiliani ER, Danielson GK, Geraci JE. Management of complications of infective endocarditis. Mayo Clin Proc 1982;57:162-70. 2. Snow RM, Cobbs CG. Treatment of complications of infective endocarditis. In: Infective endocarditis. Kaye D, ed. Baltimore: University Park Press 1976:213-27. 3. Weinstein L, Schlesinger JJ. Pathoanatomic, pathophysio-
Septic embolism complicating infective endocarditis 485
logic and clinical correlations in endocarditis. N Engl J Med 1974;291:112-26. 4. Nakayma KD, O'Neill JA, Wagner H, Cooper A, Dean RH. Management of vascular complications of bacterial endocarditis. J Pediat Surg 1986;21:636-9. 5. Pmitt AA, Rubin RH, Karchmer AW, et al. Neurologic complications of bacterial endocarditis. Medicine 1978;57: 329-43. 6. Haft JI, Altieri J, Smith LG, Herskowitz M. Computed tomography of the abdomen in the diagnosis of splenic emboli. Arch Intern Med 1988;148:193-7. 7. Balcar I, Seltzer SE, Davis S, et al. CT patterns of splenic infarction: a clinical and experimental study. Radiology 1984;151:723-9. 8. Dean RH, Waterhouse G, Meacham PW, Weaver FA, O'Neill JA. Mycotic embolism and cmbolomycotic aneurysms neglected lessons of the past. Ann Surg 1986;204:300-7. 9. Freischlag JA, Asbun HA, Sedwitz MM, Hye RJ, Sise M, Stabile BE. Septic peripheral embolization from bacterial and fungal endocarditis. Ann Vase Surg 1989;3:318-23. 10. Wilson WR, Danielson GK, Giuliani ER, et al. Cardiac valve replacement in congestive heart failure due to infective endocarditis. Mayo Clin Proc 1979;54:223-7. 11. Elliot JP, Hageman JH, Szilagyi E, et al. Arterial embolization: problems of source, multiplicity, recurrence and delayed treatment. Surgery 1980;88:833-45. 12. SymbasPN, Vlasis SE, Zacharopoutos L, et al. Immediate and long-term outlook for valve replacement in acute bacterial endocarditis. Ann Surg 1982;195:721-5. 13. D'Agostino RS, Miller DC, Stinson EB, et al. Valve replacement in patients with native valve endocarditis: what really determines operative outcome? Ann Thorac Surg 1985;40: 429-37. 14. Stinson EB. Surgical treatment of infective endocarditis. Prog Cardiovasc Dis 1979;22:145-54. 15. Karchner AW. Active infective endocarditis: when to operate. J Cardiovasc Med 1981;6:1015-31. 16. DiNubile MJ. Surgery in active endocarditis. Ann Int Med 1982;96:650-9. Submitted Feb. 2, 1991; accepted Apr. 3, 1991.
DISCUSSION
Dr. Phillip M. Levha (Los Angeles, Calif.). I congratulate the authors on a comprehensive review of a difficult management problem that requires the coordinated expertise of several specialties: cardiology, cardiac surgery, infectious disease, and vascular surgery. Whereas peripheral emboli emanating from postmyocardial infarction mural/thrombus or from rheumatic valvular disease is an almost weekly occurrence for us, our experience with IE with peripheral emboli is primarily limited to the more elderly patient with an infected valvular prosthesis.
Fortunately, the intravenous drug abuser who enters the hospital in critical condition from a cardiac standpoint and then suffers an embolic ischemic event is an uncommon occurrence at Cedars Sinai Hospital. This paper clearly demonstrates that treatment of these patients must be individualized depending on the severity of their limb ischemia and their overall cardiac status. This series was derived from 102 patients who underwent valve replacement for IE. Thirty percent of the group had septic emboli, but the paper limits its investigation to 10% of that population, eliminating the patients with
486
IGtts, Bongard, and IGein
cerebral and visceral emboli. I believe the paper would have been more informative for us if placed in the context of all patients with IE and emboli. Specifically, my first question is were the organisms associated with emboli any different from those encountered in the group as a whole? Although seven of the 10 patients examined were intravenous drug abusers, two of the three remaining patients were not. I am referring to the young woman in whom IE developed after an abortion and the young man in whom IE developed, after dental work. They represent a fascinating opportunity to observe the evolution of IE. What was the time interval from their infectious problem to the development of clinical endocarditis? Of the 10 patients examined, all but two were admitted with limb ischemia before valve replacement. Were the two remaining patients in whom peripheral emboli developed after valve replacement on any anticoagulant regimen before their peripheral emboli? If a successful outcome is to be achieved in the management of an acutely and severely ischemic limb, expeditious restoration of circulation is of paramount importance. Therefore I question the need for angiography in all patients. Not only is the time delay involved detrimental, but the diagnosis in a 20-year-old drug abuser who is admitted with fever, ischemic pain in one leg, absent ipsilateral pulses, and normal contralateral pulses can be established on clinical grounds. We would have relied more heavily on duplex scanning with color flow because it can be done quickly and clearly demonstrates the level of the occlusion. Finally, the authors discuss the development of mycotic aneurysm formation in the patients citing a 20% occurrence. These data are not found elsewhere in the paper, and I wonder where these aneurysms occurred, visceral or peripheral, and did they require any further treatment? Again, I commend the authors on an excellent presentation and an outline for management of a very difficult problem. Dr. Kitts. Thank you Dr. Levin for your comments and for reemphasizing some of the important issues this review sought to address. The first of these concerns which patients with IE and septic embolism should undergo angiography. There are two major indications for arteriography in patients with peripheral vascular emboli. First, one needs to document the level and extent of arterial occlusion to determine before operation the optimal approach to embolectomy and to predict the anticipated success of that procedure. Second, although all of our angiographically identified lesions were symptomatic, the possibility of additional asymptomatic vascular lesions, whether emboli or aneurysms, has been demonstrated in other reviews in the literature. The natural history of these asymptomatic lesions has not been documented, but progression from an occult septic embolism to a symptomatic mycotic aneurysm seems a real possibility. Knowing about these lesions ahead of time might avoid complications. With this in mind, we
Journal of VASCULAR SURGERY
have recommended that all patients with symptomatic peripheral vascular emboli considered for vascular procedures undergo angiography, and that it be used judiciously in other patients who may not require intervention. The role of duplex scanning in defining these lesions is currently being developed. Your second question, Dr. Levin, concerned the possibility of a relationship between the type of bacteria responsible for the valvular lesion and the propensity for embolization. The literature has certainly demonstrated that large, bulky, friable valvular lesions, particularly those associated with Staphylococcus aureus and fungal endocarditis have a propensity for embolization. A history of recent intravenous drug abuse was the leading risk factor for all our patients with IE considered for CVR. Streptococcusviridans was the causative organism in 70% of patients who had emboli and similarly was responsible in just over 60% of the remaining 71 patients with endocarditis without emboli. So, no, we were not able to demonstrate that the causative organism itself had as significant an impact as in previous reviews. Dr. Wiley F. Barker (Los Angeles, Calif.). I wrote a paper almost 40 years ago about mycotic aneurysms at a time when our experience was really before valve replacement of any significance, certainly not with infected valves, and before we really had even adequate antibiotic therapy. These patients, therefore, had quite a different spectrum of problems. It seemed to me we saw far fewer major infected emboli producing major ischemia. Rather, we saw more patients with lesions that resulted from smaller platelet-infected emboli lodging in smaller vessels and thus producing infected aneurysms without critical distal ischemia. From my memories of that paper I believe the most common vessel sites involved in those days included the superior mesenteric system and for some reason that is not clear, the deep femoral system. It certainly represents a change in the variety of diseases after the change in the spectrum of management. Dr. Albert D. Hall (Greenbrae, Calif.). I am curious about the origin of the endocarditis in your series. I had the impression that the most common cause was intravenous drug abuse, and then you had the dental case and the septic abortion case. Were there cases of visceral malignancy in your series of patients who had what appeared to be idiopathic endocarditis who then were discovered to have visceral malignancies such as colon or esophagus or stomach? Dr. Kitts. No. In this review 10 cases, and the cause of endocarditis was identified in every case, seven of the 10 patients had a history of recent intravenous drug abuse. One had undergone a prior prosthetic mitral valve replacement. One had had a history of recent dental extraction, and there was one case that followed therapeutic abortion. Dr. Klein. The only piece of information I would like to bring to the attention of the members of the Society relates to technical judgment if you happen to encounter a patient with a septic embolus. I would point out that the
Volume 14 Number 4 October 1991
arteritis that results from a septic embolus is exceedingly intense. Careful thought must be given before making arterial incisions because of the destructive nature of the bacterial innoculum lodged next to the arterial wall. For example, a septic popliteal embolectomy through a popliteal arteriotomy could result in fragmented artery, which would be difficult, if not impossible, to close. Because the local arterial wall essentially degenerates, remote arterial incisions for septic embolus retrieval are recommended. Dr. Wesley S. Moore (Los Angeles, Calif.). In the current era in which we are working, where the use of thrombolytic therapy is being applied to nonseptic thromboembolism, is there a role for lytic therapy in the case of septic embolism? At first it sounds like a hair-brained idea in which you are going to dissolve, clot and disseminate bacteria. But on the other hand, these patients are heavily covered with antibiotics and the dissolution ofthrombus not only at the embolic site but in other areas of the valve itself may, in fact, improve the clearance of the septic process. Are you aware of any experience in this area and would you like to speculate as to whether or not you would be willing to try it? Dr. Klein. I am aware of no published experience, and I have no experience in applying lytic therapy for a septic embolus. The problem, as Dr. Kitts alluded to, is the frequent multiplicity of septic emboli plus the cardiac valve vegetations. If dislodgement of infected material from the cardiac valve resulted while an attempt was being made to treat a peripheral focus, the possibility of initiating a remote infected region, such as a brain abscess, is likely.
Septic embolism complicating infective endocarditis 487
The second issue concerning use of lyric therapy is in making sure that one does not have a previously unrecognized cranial embolus. Several cases, not presented in this series, occurred where an asymptomatic mycotic cerebral aneurysm was not detected. The patient went on to cardiopulmonary bypass and had a devastating intracranial hemorrhage when heparinized. Similar consequences might result if lytic therapy were used without knowledge of the status of the intracranial circulation. Because of the multiplicity of septic emboli, I would recommend avoiding lytic therapy in the management of a septic embolus at the present time. Dr. Douglas Brownell (Orange, Calif.). I have one question regarding the arteriogram. You had such good results with your four patients with noncritical ischemia who were treated medically and were discharged from the hospital with no sequela. If during the arteriography you found a contralateral hypogastric or deep femoral artery embolus that was of substantial size, but obviously not critical in terms ofischemia, would you consider removing that embolus, especially in the deep femoral artery, in light of the good success with the medical treatment of the smaller more peripheral emboli? Dr. Kitts. Your point is well taken. I was quite surprised at how these patients responded to medical therapy, and the decision whether to subject that patient to embolectomy really has to be based on the degree of limb ischemia. Clearly if these patients do not have significant ischemia, they do have coexistent severe medical problems, and the literature demonstrates that these patients will frequently respond to conservative medical therapy. I think that probably should be tried first.
Torrance, Calif. Embolic phenomena in patients with infective endocarditis may complicate the placement of a cardiac vah~lar prosthesis. To evaluate the vascular consequences of these emboli, a 15-year review of 102 patients undergoing valve replacement for proven infective endocarditis was undertaken. Thirty-one patients with 36 episodes of septic embolization were identified. Ten of these were separate extremity occlusive events. All patients with extremity emboli were admitted with pain; four had limb-threatening emboli. All patients grew gram-positive bacteria from their blood except a single Candida albicans isolate. Appropriate antimicrobial therapy was used in all patients. Angiography confirmed the diagnosis in 11 of 12 patients. Embolic targets included the lower extremities in all except a single instance. Four patients had multiple emboli. All but one of the vascular procedures were carried out subsequent to or simultaneously with cardiac valve replacement. Initial operative management included embolectomy (4) and primary amputation (2). Two delayed procedures were required. One patient died. Four patients had limited ischemia that resolved with antibiotics and anticoagulation. This report suggests that infective endocarditis requiring valvaxlar replacement is associated with embolization in one third of patients. The presentation of peripheral vascular emboli is that of acute extremity ischemia. The diagnosis should be confirmed by angiography to rule out the possibility of multiple emboli. When possible, valve replacement should precede peripheral vascular management, which may include operative or medical components as dictated by the extent of limb ischemia. (J VASe SuRe 1991;14:480-7.)
Vascular surgeons are increasingly being called on to recognize the morbidity imposed by coexistent cardiac disease in patients seeking treatment for extremity ischemia. In this regard, few patients will present as challenging a problem as the patient with infective endocarditis (IE) and extremity ischemia caused by septic embolism. Management may be complicated by poor cardiac status as well as the effects of additional emboli to the cerebral, visceral, and other peripheral vascular beds. 1-4 Patients with IE referred for cardiac valve replacement (CVR) are a source of such management problems because many have experienced multiple embolic events as well as congestive heart failure. The purpose of this report is to delineate the key factors leading to From the Departmentof Surgery,Harbor-UCLAMedicalCenter, Torrance. Presented at the Sixth AnnualMeeting of the Western Vascular Society, Palm Springs, Calif.,Jan. 13-16, 1991. Reprint requests: StanleyR. Klein,MD, Department of Surgery, Harbor-UCLA MedicalCenter, 1000 W. Carson St., Box 15, Torrance CA 90509. 24/6/29914 480
successful outcome and minimal morbidity in the management of these patients with IE and peripheral septic embolism. METHODS
Cardiac valve replacement for proven IE was performed in 102 patients at Harbor-UCLA Medical Center between January 1974 and December 1989. Thirty-one of these patients (30%) had septic embolic events: nine patients had cerebral emboli, six had visceral emboli, six had isolated pulmonary emboli, and 10 patients had major peripheral vascular embolization. The characteristics and management of the latter 10 patients with primarily peripheral vascular emboli form the basis of this report. PATIENTS The 10 patients in this report include six men and four women with ages ranging from 20 to 44 years at the time of initial presentation. Individual predisposing conditions, clinical course, and management are outlined in Table I. A history of recent intravenous drug abuse was the predominant risk factor in
Volume 14 Number 4 October 1991
Septic embolism complicating infective endocarditis
481
Table I. Individual predisposing conditions, clinical course, patient management Patient (age)
Risk factor
1. L.G. (31)
IVDA, bicuspid AV
2. M.D. (24)
IVDA
3. C.T. (24)
prior MVR
4. M.M. (25)
IVDA
5. D.R. (32)
IVDA
6. A.L. (44)
IVDA
7. R.M. (43)
IVDA
8. R.J. (23)
Septic TAB
9. R.B. (26)
Dental work
10. A.B. (20)
IVDA
Symptoms on admission
Valve
Embolism
Culture
Management
Fever, CHF, bilateral ischemic leg pain Fever, CHF, abdominal pain, ischemic right leg pain Fever, ischemic left foot pain
Aortic
Bilateral external lilac arteries
Aortic
Right SFA
Mitral
Left femoral/popliteal embotectomy
Good
Fever, LUQ abdominal pain, ischemic left foot pain Fever, CHF, hypotension, ischemic left foot pain Fever, CHF, hypotension Fever, CHF, left facial paralysis, ischemic left leg pain Fever, hypotension, LU, RL extremity paresthesias Fever, ischemic left leg pain Fever, CHF, weight loss
Aortic/mitral
Left common s. epi femoral artery, left popliteal artery, right renal artery Left popliteal S. viridans artery, spleen, right MCA
AVR/MVR, splenectomy, delayed femoral distal bypass
Good
Aortic mitral
Left tibial artery
S. faecalis
AVR/MVR, delayed transmetatarsal amputation
Good
Aortic
Spleen left popliteal artery Left common iliac artery, right MCA aneurysm
AVR, splenecS. viridans tomy, delayed C. albicans left BKA AVR, Coumadin S. viridans
Died
Aortic
Candida AVR parapsilosis Bilateral femoral/embolectomy Streptococcus AVR, right SFA viridans embolectomy
Outcome
Good
Good
Good
Mitral
Left brachial Group B artery, right streptopopliteal cocci artery, spleen
MVR, Coumadin
Good
Aortic
Left tibial artery
S. viridans
AVR, Coumadin, antibiotics
Good
Mitral
Right cornmon lilac artery, left superficial femoral artery
S. viridans
MVR, Coumadin
Good
IVDA, Intravenous drug abuse; CHF, congestive heart failure; LV, left upper; AVR, aortic valve replacement; LUQ, left upper quadrant; TAB, therapeutic abortion; RL, right lower; SFA, superficial femoral artery; MCA, middle cerebral artery; BKA, below-knee amputation.
this group, occurring in 7 of 10 patients (70%). Exceptions included a 24-year-old woman (case 3) with porcine mitral valve replacement 1 year previously, a 23-year-old woman (case 8) with septic endometritis 5 days after therapeutic abortion, and a 26-year-old man (case 9) with recent dental extractions.
CLINICAL PRESENTATION All patients were admitted with fever, and all but two (80%) had ischemic extremity pain. Embolic occlusive events after CVR developed in both patients without ischemia on admission (cases 6 and 8). Congestive heart failure with pulmonary edema was present in 50% of patients, and three patients
482
Journal of VASCULAR SURGERY
IGtts, Bongard, and IGein
Table II. Site of embolism (10 patients) Site (artery)
No. of events
Common iliac External iliac Common femoral Superficial femoral Popliteal Tibial Brachial
2 2 1 2 4 2 1
Total
14
(30%) had hypotension with systolic blood pressure less than 90 mm Hg on admission. In addition, one patient (case 7), had a dense neurologic deficit as a result of cerebral aneurysm, and another had abdominal pain caused by splenic embolism. Overall, 50% of patients were demonstrated to have remote multiorgan emboli in addition to their primary vascular event. Three patients had concomitant splenic emboli (cases 4, 6, and 8), two patients had cerebral lesions (cases 4 and 7), and a single patient (case 3), had a right renal embolus. BACTERIOLOGY The source of septic embolism was an isolated aortic valvular lesion in five patients, both aortic and mitral involvement in two patients, and isolated mitral valvular infection in the remaining three patients. All patients had at least two positive blood cultures on separate occasions. Gram stain and culture of valvular vegetations demonstrated the offending organism in 100% of cases. Eight of the nine patients with bacterial endocarditis (89%) had streptococcal infections, the only exception being the prosthetic valve infection caused by Staphylococcus epidermidis. A single patient (case 1), was admitted with fungal endocarditis caused by Candida parapsi-
losis. PRESENTATION OF LESIONS
Eight o f the 10 patients in this series were symptomatic from major peripheral vascular emboli at the time of admission. Septic emboli developed in the two additional patients and they became symptomatic shortly after CVR (cases 6 and 10). Initial vascular examination demonstrated a pulse deficit distal to ithe embolic site in 9 of the 10 patients. Neurologic impairment manifested as extremity sensory deficit was present in four patients, and four patients had overt limb threat (cases 1 to 4). All patients underwent angiography before any vascular procedure. Fourteen embolic events were
identified in the 10 patients. Embolic targets were as outlined in Table II. Included were four patients who had multiple simultaneous peripheral emboli. The patients with multiple emboli had lesions of the bilateral external iliac arteries (1), common iliac and contralateral common femoral arteries (1), common femoral and ipsilateral popliteal arteries (1), and popliteal and contralateral brachial arteries (1). All of the angiographically identified lesions were symptomatic. It is of note that there were no asymptomatic lesions detected. MANAGEMENT Treatment of embolic episodes varied considerably between patients and was dependent on cardiac status, magnitude of limb ischemia, and severity of concurrent illness caused by other remote emboli. Patients could be grouped into (1) those requiring emergent CVR and simultaneous embolectomy, (2) patients requiring emergent embolectomy whose cardiac status was sufficiently stable to tolerate a noncardiac procedure, and (3) patients in whom the degree of limb ischemia did not warrant the risk of emergency intervention. EMERGENCY EMBOLECTOMY
Four patients underwent emergency embolectomy. In two patients (cases 1 and 2) this was performed simultaneously with CVR. Both patients were admitted with florid congestive heart failure as a result of aortic regurgitation in addition to limb ischemia. Cardiac status would have prohibited initial embolectomy. Concomitant CVR and embolectomy resulted in excellent outcome in both cases with resolution of ischemia and complete recovery before discharge from the hospital. The third emergency embolectomy (case 3) was performed before CVR for left lower extremity limb threat caused by common femoral artery and popliteal artery embolic occlusion. Intraoperative invasive hemodynamic monitoring of cardiac function allowed for uncomplicated general anesthesia and embolectomy despite mitral insufficiency. The concomitant right renal embolism remained asymptomatic, and the patient had an uneventful complete recovery after subsequent CVR. The final emergency embolectomy was performed in case 4, a 25-year-old man admitted with high fever, left upper quadrant abdominal pain, and an ischemic left lower extremity. Abdominal CT scanning demonstrated a splenic abscess, and angiography revealed isolated left popliteal occlusion. This patient's cardiac status allowed for simultaneous
Volttme 14 Number 4 October 1991
Septic embolism complicating infective endocarditis 483
splenectomy and left popliteal embolectomy resulting in a viable extremity. Subsequent CVR was complicated by left hemiparesis as a result of right middle cerebral embolism. After a prolonged recovery he had left leg claudication necessitating femoral posterior tibial bypass for chronic popliteal occlusion. He was eventually discharged from the hospital with no symptoms and fully recovered 5 months after initial admission. AMPUTATION Two patients were treated with amputation as the initial procedure for peripheral vascular embolism (cases 5 and 6). The first of these, a 32-year-old woman was admitted with congestive heart failure, left foot pain, and absent pedal pulses and required emergent aortic valve replacement/mitral valve replacement for hemodynamic instability. Intraoperatire angiography demonstrated the crural vessels to be patent to the ankle. Despite this, dry gangrene of the left forefoot developed after operation because of distal tibial emboli. She underwent transmetatarsal amputation 6 weeks after CVR and was discharged from the hospital able to walk with a cane. The remaining amputation was performed in a 44-yearold man admitted with congestive heart failure and sepsis who also required emergent aortic valve replacement for progressive deterioration (case 6). His postoperative course was complicated by splenic abscess requiring splenectomy and later multisystem organ failure caused by Candida sepsis. Three weeks after CVR his left foot became pulseless, but ongoing hemodynamic instability precluded intervention. He subsequently underwent a left below-knee amputation as a last resort but died shortly thereafter of progressive multisystem organ failure. NONOPERATIVE MANAGEMENT Nonoperative management consisting of antibiotics and anticoagulation was used in the remaining four patients. In three of the four (cases 7, 8, and 10) the extent of fimb ischemia did not warrant aggressive intervention given the coexistent cardiac insufficiency. One of these patients (case 7) was admitted with congestive heart failure and right middle cerebral stroke in addition to left lower extremity pain. Angiography demonstrated a left common iliac artery occlusion with reconstitution of the common femoral artery. Despite the impressive arteriogram (Fig. 1), his ischemic extremity symptoms completely resolved with antibiotics and anticoagulation before CVR. At follow-up after discharge from the
Fig. 1. Arteriogram shows left common iliac occlusion.
hospital he remained asymptomatic and had regained full femoral and distal pulses. Two additional patients were admitted with ischemic extremity pain without limb threat (cases 8 and 9). In both of these antibiotics and anticoagulation again resulted in complete resolution of symptoms during the period of treatment before CVR. The final patient treated nonoperatively had lower extremity pain 2 days after emergent mitral valve replacement for IE and progressive deterioration. Angiography demonstrated right common iliac artery and left superficial femoral artery occlusion with distal reconstitution. Embolectomy was recommended, but the patient refused operation. He was treated with antibiotics, and sodium warfarin (Coumadin) was continued. Two weeks after this event, he left the hospital against advice and was asymptomatic with exertion. DISCUSSION Excluding congestive heart failure, arterial embolization of infected debris and its sequelae is the most common complication of IE. This event occurs in
484 IGtts, Bongard, and IGein
approximately 30% of patients with acute IE in published series.l-3 We found a similar incidence of septic embolism in our patients with acute IE referred for CVR. Thirty-one of the 102 patients (30%) undergoing CVR for IE had major documented embolic events. The manifestation of such an event depends ultimately on the destination of the embolus. Although pulmonary, splenic, and renal emboli are frequently well tolerated, a single embolus to the cerebral, coronary, or peripheral vascular bed can permanently disable a previously fully functional patient. Unlike the cerebral and coronary circulations, the peripheral vasculature is readily accessible and provides the opportunity to potentially reverse organ ischemia and tissue death. Management may be complicated, however, by coexistent cardiac dysfunction and the added morbidity of simultaneous emboli to the cerebral, visceral, and other peripheral vascular targets. Fifty percent of patients in this series were admitted in congestive heart failure. Half again had multiorgan emboli in addition to limb ischemia, including three splenic, one renal, and two cerebral lesions. This tendency for septic embolization to be recurrent and widespread has been noted previously. Six of eight patients (75%) reported by Nakayama et al.4 and 50% of patients in the series by Pruitt et al.5 suffered multiple systemic emboli. This has serious implications if complications from unrecognized cerebral and visceral emboli are to be avoided. The single death in this series occurred in a patient who underwent emergent aortic valve replacement and subsequently required splenectomy for ongoing sepsis caused by splenic abscess not recognized before operation. Although unavoidable in this instance, a valuable lesson was reiterated. The sensitivity of CT scanning in detecting splenic emboli is well documented, and many authorities believe that this should be routinely performed in patients with IE to avoid such complications.6"7 Multiple peripheral vascular emboli and "embolomycotic" aneurysms may also complicate the course of IE. Seven of nine patients in the series by Dean et al.8 had multiple vascular emboli, and 14 aneurysms were discovered in these nine patients. In the present series, angiography detected multiple vascular emboli in four patients (40%), and embolomycotic aneurysms were detected in two patients (20%). Although all embolic lesions in this series were symptomatic, the propensity for septic emboli and particularly aneurysms to remain silent has been stressed by Nakayama et al.,4 Dean et al.,8 and Freischlag et al.9
Journal of VASCULAR SURGERY
Angiography to date represents the only means of recognizing and treating these potentially silent lesions and has become an integral part of the vascular workup in patients with IE and septic embolism. 4"8"9 The noninvasive vascular laboratory has little experience in this regard. Although embolectomy may be one of the less challenging procedures undertaken by vascular surgeons, treatment of the patient with limb ischemia and IE promises to be a complex problem. Even in the face of limb threat, cardiac performance and hemodynamic stability must remain priorities. Patients admitted while in congestive heart failure caused by valvular incompetence will poorly tolerate any form of anesthesia associated with a lengthy vascular procedure. 1° It is important for the vascular surgeon to know that the endarteritis that rapidly results from the infected embolus frequently destroys the integrity of the arterial wall. Thus what may begin as a "simple embolectomy" often escalates into a more complicated revascularization requiring a much longer period of time. It is also important to recognize that valuable time may be lost during attempted diuresis and afterload reduction to improve cardiac performance. The direct correlation between preoperative delay and subsequent amputation was emphasized by Elliot et al.ll in their review of 225 patients with nonseptic arterial emboli. After 8 hours the amputation rate was 50% higher, and after 24 hours it had doubled. 11 In the setting of cardiac decomposition caused by IE and limb threat as a result of concomitant septic embolism, serious consideration should be given to simukaneous emergent CVR and embolectomy. Cardiac surgeons may be reluctant to proceed with emergent CVR in the presence of acute IE, citing the potential for increased prosthetic valve infection in this setting. The incidence of prosthetic valvular reinfection in acute IE is less than 10%. 12-14 Furthermore, the benefit of limb salvage with simultaneous eradication of the source of additional emboli may outweigh this risk. In those instances in which limb threat is present and CVR is clearly not indicated, invasive intraoperative hemodynamic monitoring and the presence of an experienced anesthetist may prevent cardiac decompensation during a procedure. Finally, the magnitude of limb ischemia may not warrant embolectomy, particularly if the patient's condition is complicated by congestive heart failure or stroke as was the case in one of our patients. Karchner, 15 DiNubile, 16 and others have demonstrated that antibiotics and often anticoagulation
Volume 14 Number 4 October 1991
were frequently sufficient to treat embolic complications o f IE. F o u r patients in this series were treated in this m a n n e r and all had complete resolution o f ischemic symptoms. CONCLUSION Septic embolization occurs in at least 30% o f patients with I E referred for C V R . A high index o f suspicion m u s t be maintained if this complication is to be detected. O n e third o f these embolic events are to the pcripheral vasculaturc. Such emboli arc frequently multiple and m a y be accompanied by mycotic aneurysms, or splenic, cerebral, and renal infarcts. A n g i o g r a p h y and abdominal C T scanning are invaluable in determining before operation the sites o f septic embolism. A low threshold for obtaining these studies is indicated. The noninvasive laboratory m i g h t provide an excellent screening modality in the future. T h e presentation is frequently that o f limb ischemia superimposed o n congestive heart failure. T h e extent o f limb ischemia and degree o f cardiac dysfunction will dictate the priority o f the problems. W h e n possible, C V R should proceed before or simultaneously with e m b o l e c t o m y to avoid the m o r b i d i t y o f e m b o l e c t o m y in a hemodynamically unstable patient.
REFERENCES
1. Wilson WR, Guiliani ER, Danielson GK, Geraci JE. Management of complications of infective endocarditis. Mayo Clin Proc 1982;57:162-70. 2. Snow RM, Cobbs CG. Treatment of complications of infective endocarditis. In: Infective endocarditis. Kaye D, ed. Baltimore: University Park Press 1976:213-27. 3. Weinstein L, Schlesinger JJ. Pathoanatomic, pathophysio-
Septic embolism complicating infective endocarditis 485
logic and clinical correlations in endocarditis. N Engl J Med 1974;291:112-26. 4. Nakayma KD, O'Neill JA, Wagner H, Cooper A, Dean RH. Management of vascular complications of bacterial endocarditis. J Pediat Surg 1986;21:636-9. 5. Pmitt AA, Rubin RH, Karchmer AW, et al. Neurologic complications of bacterial endocarditis. Medicine 1978;57: 329-43. 6. Haft JI, Altieri J, Smith LG, Herskowitz M. Computed tomography of the abdomen in the diagnosis of splenic emboli. Arch Intern Med 1988;148:193-7. 7. Balcar I, Seltzer SE, Davis S, et al. CT patterns of splenic infarction: a clinical and experimental study. Radiology 1984;151:723-9. 8. Dean RH, Waterhouse G, Meacham PW, Weaver FA, O'Neill JA. Mycotic embolism and cmbolomycotic aneurysms neglected lessons of the past. Ann Surg 1986;204:300-7. 9. Freischlag JA, Asbun HA, Sedwitz MM, Hye RJ, Sise M, Stabile BE. Septic peripheral embolization from bacterial and fungal endocarditis. Ann Vase Surg 1989;3:318-23. 10. Wilson WR, Danielson GK, Giuliani ER, et al. Cardiac valve replacement in congestive heart failure due to infective endocarditis. Mayo Clin Proc 1979;54:223-7. 11. Elliot JP, Hageman JH, Szilagyi E, et al. Arterial embolization: problems of source, multiplicity, recurrence and delayed treatment. Surgery 1980;88:833-45. 12. SymbasPN, Vlasis SE, Zacharopoutos L, et al. Immediate and long-term outlook for valve replacement in acute bacterial endocarditis. Ann Surg 1982;195:721-5. 13. D'Agostino RS, Miller DC, Stinson EB, et al. Valve replacement in patients with native valve endocarditis: what really determines operative outcome? Ann Thorac Surg 1985;40: 429-37. 14. Stinson EB. Surgical treatment of infective endocarditis. Prog Cardiovasc Dis 1979;22:145-54. 15. Karchner AW. Active infective endocarditis: when to operate. J Cardiovasc Med 1981;6:1015-31. 16. DiNubile MJ. Surgery in active endocarditis. Ann Int Med 1982;96:650-9. Submitted Feb. 2, 1991; accepted Apr. 3, 1991.
DISCUSSION
Dr. Phillip M. Levha (Los Angeles, Calif.). I congratulate the authors on a comprehensive review of a difficult management problem that requires the coordinated expertise of several specialties: cardiology, cardiac surgery, infectious disease, and vascular surgery. Whereas peripheral emboli emanating from postmyocardial infarction mural/thrombus or from rheumatic valvular disease is an almost weekly occurrence for us, our experience with IE with peripheral emboli is primarily limited to the more elderly patient with an infected valvular prosthesis.
Fortunately, the intravenous drug abuser who enters the hospital in critical condition from a cardiac standpoint and then suffers an embolic ischemic event is an uncommon occurrence at Cedars Sinai Hospital. This paper clearly demonstrates that treatment of these patients must be individualized depending on the severity of their limb ischemia and their overall cardiac status. This series was derived from 102 patients who underwent valve replacement for IE. Thirty percent of the group had septic emboli, but the paper limits its investigation to 10% of that population, eliminating the patients with
486
IGtts, Bongard, and IGein
cerebral and visceral emboli. I believe the paper would have been more informative for us if placed in the context of all patients with IE and emboli. Specifically, my first question is were the organisms associated with emboli any different from those encountered in the group as a whole? Although seven of the 10 patients examined were intravenous drug abusers, two of the three remaining patients were not. I am referring to the young woman in whom IE developed after an abortion and the young man in whom IE developed, after dental work. They represent a fascinating opportunity to observe the evolution of IE. What was the time interval from their infectious problem to the development of clinical endocarditis? Of the 10 patients examined, all but two were admitted with limb ischemia before valve replacement. Were the two remaining patients in whom peripheral emboli developed after valve replacement on any anticoagulant regimen before their peripheral emboli? If a successful outcome is to be achieved in the management of an acutely and severely ischemic limb, expeditious restoration of circulation is of paramount importance. Therefore I question the need for angiography in all patients. Not only is the time delay involved detrimental, but the diagnosis in a 20-year-old drug abuser who is admitted with fever, ischemic pain in one leg, absent ipsilateral pulses, and normal contralateral pulses can be established on clinical grounds. We would have relied more heavily on duplex scanning with color flow because it can be done quickly and clearly demonstrates the level of the occlusion. Finally, the authors discuss the development of mycotic aneurysm formation in the patients citing a 20% occurrence. These data are not found elsewhere in the paper, and I wonder where these aneurysms occurred, visceral or peripheral, and did they require any further treatment? Again, I commend the authors on an excellent presentation and an outline for management of a very difficult problem. Dr. Kitts. Thank you Dr. Levin for your comments and for reemphasizing some of the important issues this review sought to address. The first of these concerns which patients with IE and septic embolism should undergo angiography. There are two major indications for arteriography in patients with peripheral vascular emboli. First, one needs to document the level and extent of arterial occlusion to determine before operation the optimal approach to embolectomy and to predict the anticipated success of that procedure. Second, although all of our angiographically identified lesions were symptomatic, the possibility of additional asymptomatic vascular lesions, whether emboli or aneurysms, has been demonstrated in other reviews in the literature. The natural history of these asymptomatic lesions has not been documented, but progression from an occult septic embolism to a symptomatic mycotic aneurysm seems a real possibility. Knowing about these lesions ahead of time might avoid complications. With this in mind, we
Journal of VASCULAR SURGERY
have recommended that all patients with symptomatic peripheral vascular emboli considered for vascular procedures undergo angiography, and that it be used judiciously in other patients who may not require intervention. The role of duplex scanning in defining these lesions is currently being developed. Your second question, Dr. Levin, concerned the possibility of a relationship between the type of bacteria responsible for the valvular lesion and the propensity for embolization. The literature has certainly demonstrated that large, bulky, friable valvular lesions, particularly those associated with Staphylococcus aureus and fungal endocarditis have a propensity for embolization. A history of recent intravenous drug abuse was the leading risk factor for all our patients with IE considered for CVR. Streptococcusviridans was the causative organism in 70% of patients who had emboli and similarly was responsible in just over 60% of the remaining 71 patients with endocarditis without emboli. So, no, we were not able to demonstrate that the causative organism itself had as significant an impact as in previous reviews. Dr. Wiley F. Barker (Los Angeles, Calif.). I wrote a paper almost 40 years ago about mycotic aneurysms at a time when our experience was really before valve replacement of any significance, certainly not with infected valves, and before we really had even adequate antibiotic therapy. These patients, therefore, had quite a different spectrum of problems. It seemed to me we saw far fewer major infected emboli producing major ischemia. Rather, we saw more patients with lesions that resulted from smaller platelet-infected emboli lodging in smaller vessels and thus producing infected aneurysms without critical distal ischemia. From my memories of that paper I believe the most common vessel sites involved in those days included the superior mesenteric system and for some reason that is not clear, the deep femoral system. It certainly represents a change in the variety of diseases after the change in the spectrum of management. Dr. Albert D. Hall (Greenbrae, Calif.). I am curious about the origin of the endocarditis in your series. I had the impression that the most common cause was intravenous drug abuse, and then you had the dental case and the septic abortion case. Were there cases of visceral malignancy in your series of patients who had what appeared to be idiopathic endocarditis who then were discovered to have visceral malignancies such as colon or esophagus or stomach? Dr. Kitts. No. In this review 10 cases, and the cause of endocarditis was identified in every case, seven of the 10 patients had a history of recent intravenous drug abuse. One had undergone a prior prosthetic mitral valve replacement. One had had a history of recent dental extraction, and there was one case that followed therapeutic abortion. Dr. Klein. The only piece of information I would like to bring to the attention of the members of the Society relates to technical judgment if you happen to encounter a patient with a septic embolus. I would point out that the
Volume 14 Number 4 October 1991
arteritis that results from a septic embolus is exceedingly intense. Careful thought must be given before making arterial incisions because of the destructive nature of the bacterial innoculum lodged next to the arterial wall. For example, a septic popliteal embolectomy through a popliteal arteriotomy could result in fragmented artery, which would be difficult, if not impossible, to close. Because the local arterial wall essentially degenerates, remote arterial incisions for septic embolus retrieval are recommended. Dr. Wesley S. Moore (Los Angeles, Calif.). In the current era in which we are working, where the use of thrombolytic therapy is being applied to nonseptic thromboembolism, is there a role for lytic therapy in the case of septic embolism? At first it sounds like a hair-brained idea in which you are going to dissolve, clot and disseminate bacteria. But on the other hand, these patients are heavily covered with antibiotics and the dissolution ofthrombus not only at the embolic site but in other areas of the valve itself may, in fact, improve the clearance of the septic process. Are you aware of any experience in this area and would you like to speculate as to whether or not you would be willing to try it? Dr. Klein. I am aware of no published experience, and I have no experience in applying lytic therapy for a septic embolus. The problem, as Dr. Kitts alluded to, is the frequent multiplicity of septic emboli plus the cardiac valve vegetations. If dislodgement of infected material from the cardiac valve resulted while an attempt was being made to treat a peripheral focus, the possibility of initiating a remote infected region, such as a brain abscess, is likely.
Septic embolism complicating infective endocarditis 487
The second issue concerning use of lyric therapy is in making sure that one does not have a previously unrecognized cranial embolus. Several cases, not presented in this series, occurred where an asymptomatic mycotic cerebral aneurysm was not detected. The patient went on to cardiopulmonary bypass and had a devastating intracranial hemorrhage when heparinized. Similar consequences might result if lytic therapy were used without knowledge of the status of the intracranial circulation. Because of the multiplicity of septic emboli, I would recommend avoiding lytic therapy in the management of a septic embolus at the present time. Dr. Douglas Brownell (Orange, Calif.). I have one question regarding the arteriogram. You had such good results with your four patients with noncritical ischemia who were treated medically and were discharged from the hospital with no sequela. If during the arteriography you found a contralateral hypogastric or deep femoral artery embolus that was of substantial size, but obviously not critical in terms ofischemia, would you consider removing that embolus, especially in the deep femoral artery, in light of the good success with the medical treatment of the smaller more peripheral emboli? Dr. Kitts. Your point is well taken. I was quite surprised at how these patients responded to medical therapy, and the decision whether to subject that patient to embolectomy really has to be based on the degree of limb ischemia. Clearly if these patients do not have significant ischemia, they do have coexistent severe medical problems, and the literature demonstrates that these patients will frequently respond to conservative medical therapy. I think that probably should be tried first.
