History and Terminology

Goals of this Discussion

Case Presentation

Epidemiology

Pathogenesis

Pathological Findings

Clinical Features

Diagnosis

Therapy

Summary and Key Points

Nonbacterial thrombotic endocarditis (NBTE) is a term first used by Gross and Friedburg¹ in 1936. Many other terms have been used to describe this entity, including thromboendocarditis², cachectic or marantic³ endocarditis, endocarditis of "terminal type"⁴, "degenerative verrucous endocarditis"⁵, and more recently, aseptic valvar thrombosis⁶. Currently, it is most commonly referred to as NBTE, which is what it will be called for this discussion.

NBTE is pathologically defined as "agglutinated blood and platelet-fibrin thrombi" that are superficially attached to valve tissue with the absence of inflammation^1-4. It was first associated with the terminal phases of many neoplastic diseases. While still most commonly found in patients with advanced malignancies, it is now known to also occur in patients with other diseases and at any stage in a disease’s time course. NBTE was largely discounted in the literature until the 1950’s when several different investigators presented evidence that it resulted in clinically important events, chiefly arterial embolization^7-8. Today, NBTE is still vastly ignored clinically and entertained mostly postmortem at clinicopathologic conferences. Perhaps now that our understanding of its pathogenesis is unfolding, effective non-invasive diagnostic tools are emerging , and potential treatment options are under development, NBTE will become a more readily recognizable and treatable phenomenon by clinicians at the bedside of living patients.

Goals of this Discussion

The goals of this paper will be to first discuss two cases of NBTE encountered within a one month period on a general medicine ward at a teaching hospital that were diagnosed antemortem. A review of the literature will then be presented outlining the epidemiology, pathogenesis, clinical features, pathologic findings, diagnosis, and therapy of NBTE. Current concepts in the understanding of the pathogenesis of NBTE, emerging diagnostic strategies, and potential therapeutic options will be emphasized. More generally, and perhaps more importantly, an attempt will be made to convince fellow physicians that NBTE is of clinical importance to them.

Most doctors do not know much about NBTE and consider it a rare, silent, and terminal process of unknown etiology that is impossible to diagnose in living patients soon enough to initiate even unproven therapy to them. Hopefully after this discussion, physicians will appreciate that NBTE is a clinically relevant process that is not so rare. In fact, if it is considered and sought out, primary care physicians, cardiologists, infectious disease specialists, neurologists, hematologists, and oncologists will encounter it as often as the pathologists do currently. It will also become evident that NBTE is often diagnosable and, in some patients, treatable.

Case #1 – RR - Presentation

RR was a previously healthy, athletic 37 year-old white male who was initially admitted to the Neurology service 9/4/99 with "mental status changes", including generalized confusion and an expressive aphasia. MRI of the brain at that time revealed a large left posterior MCA infarct and several smaller infarcts involving both cerebral hemispheres and the cerebellum (see figure #1). The ensuing work-up included a transesophageal echocardiogram (TEE) which showed a small (» 2mm) vegetation on the atrial surface of the mitral valve. He denied any symptoms preceding his aphasia, specifically denying fever, chills, or weight loss. He had no overt stigmata of systemic emboli on initial exam, and laboratory studies were only remarkable for a mild leukocytosis (13,200). Of note, he had been to the dentist three weeks prior for routine cleaning. Four sets of blood cultures were obtained prior to the initiation of broad spectrum antibiotics for presumed infective endocarditis with septic emboli. His initial hospital stay was uneventful, with no fevers or further clinically evident emboli. He had gradual improvement in his confusion, and after nine negative blood cultures, he was discharged on 9/11/99 to complete a four week course of ceftriaxone and doxycycline per the recommendations of the infectious disease consultant for "culture negative endocarditis."

On 9/15/99 RR went to his local emergency department with worsening of his confusion and nausea and vomiting. Cranial CT at that time showed no new infarcts. His symptoms rapidly worsened at home, and he was readmitted on 9/17/99 for a marked decline in his mental status, increasing abdominal pain, and nausea. His family denied all other complaints on an extensive review of systems.

Figure 1. MRI of the brain in a patient with nonbacterial thrombotic endocarditis 
(RR - case #1) demonstrating bilateral embolic events.

Past Medical History - negative until CVA on 9/4/99

Medications - Zantac, Aspirin, Ceftriaxone, Doxycycline

Social History – smoked ½ pack/day of cigarettes for 12 years, occasional alcohol, no IV drug abuse or HIV risk factors, worked as a firefighter and Army Reservist

Family History – father previously treated for CLL

Physical Exam –afebrile, Pulse=90, BP=128/72

Laboratory Data – WBC=20,800 with 81% neutrophils, Hgb=13.6, plts=170,000

Studies – 9/17 – RUQ Ultrasound – multiple gallstones but no acute cholycystitis

9/18 – CT scan of abdomen/pelvis (figure 2) – multiple low density liver lesions up to 1.7cm, normal biliary tree, 2.7cm lesion in the head of the pancreas, right renal infarct, small retroperitoneal lymph nodes.  Read as "consistent with septic emboli, but could represent pancreatic cancer with liver metastases"

Figure 2.  Abdominal CT scan of patient #1 (RR) demonstrating low density liver lesions,
typical wedge-shaped renal infarct, and pancreatic lesion.

9/19 - Repeat TEE – small flat protuberance seen on atrial surface of the mitral valve, preserved LV function with mild mitral regurgitation

Hospital Course

RR was transferred to the general medicine service on 9/19/99 for further work-up. He was continued on antibiotics for "culture-negative" endocarditis and blood cultures were obtained daily. On 9/19 he underwent a repeat TEE which again revealed an extremely small vegetation in the same place on the atrial surface of the mitral valve. The CT surgery consultants wanted a more complete investigation before operative options were to be discussed.

The general medicine team and all of the consulting specialists felt that a biopsy specimen would be needed to ultimately answer the etiologic question in this young patient. In addition to culture-negative endocarditis with septic emboli and nonbacterial thrombotic endocarditis with an underlying malignancy, such entities as atrial mxyoma and diffuse vasculitis were considered. On 9/21/99 the patient underwent a CT-guided biopsy of one of the more peripheral liver lesions which was preliminarily "worrisome for malignancy, type unknown." For a definitive diagnosis, RR underwent esophagogastroduodenoscopy with endoscopic ultrasound which clearly revealed a large, solid pancreatic mass which was eroding into the portal vein (figure 3). A transduodenal biopsy specimen was consistent with pancreatic carcinoma.

Figure 3.  Endoscopic ultrasound in case #1 (RR) revealing large
pancreatic mass eroding into the portal vein (PV).

Oncology was consulted and offered RR palliative Gemzar. The patient was adamant about going home quickly, so arrangements were made for RR to follow-up with Oncology as an out-patient on 9/25/99. He and his family were to meet with Hospice at that time. On 9/23/99 RR was discharged to home with his family in an ambulatory condition. His abdominal pain was controlled with narcotics. The potential role for anti-coagulation was discussed with RR and his family, but given his terminal diagnosis, the rapidity of progression, and the lack of absolute indication, they refused this option.

Case #2 – KF – Presentation

KF was a 56 year-old white female with a long-standing history of secondary progressive multiple sclerosis (MS). She periodically had "flares" of her disease that were characterized by increased generalized weakness and confusion and which seemed to partially respond to a short course of intravenous steroids. She had received this therapy for similar episodes in June, July, and late September of 1999. KF was then admitted to the Neurology service at North Carolina Baptist Hospital on 10/20/99 with new-onset seizures, anemia, and leukocytosis in addition to anorexia and generalized weakness.

Past Medical History – Secondary progressive MS > 10 years, previous hysterectomy

Medications – MS meds including Betaseron, Aricept, INH, Zaneflex, Lithium, B12, Prozac, and Calcium Carbonate, plus Ibuprofen and Tylenol

Family History – Mother died of breast cancer and a she has a first cousin with multiple sclerosis

Social History – Smoked 1 pack/day for 18 years, no alcohol or intravenous drugs

Physical Exam – Temp=97.5, BP=162/92, pulse=96

Laboratory Data – WBC=28,000 with 79% segs and 2% bands, Hgb=8.0, plts=250,000

Initial Studies – Uninfused CT of Head – mild leukoariosis as before but no new intracracnial abnormality

Hospital Course

KF was quite lethargic for several hours. Upon finally awakening, she complained of severe abdominal pain that was difficult to localize, but seemed greatest in the right-upper quadrant. Further questioning revealed that her chronic abdominal pain had changed in character over the preceding day. An abdominal CT scan (figure 4) was performed on 10/21/99 which revealed numerous areas of infarction in the spleen, kidneys, and liver consistent with emboli, in addition to a large gallbladder with mucosal sloughing that the general surgery consultants wanted to treat medically initially.

A TEE done on 10/21/99 revealed multiple large densities on the atrial surface of all mitral valve leaflets and normal cardiac function. She was transferred to a general medicine service on 10/21/99 and after 4 blood cultures were collected she was started on broad-spectrum antibiotics for presumed infective endocarditis.

Figure 4.  Abdominal CT scan of patient #2 (KF) demonstrating large
gallbladder, hypodense infarcts of the liver and spleen, and wedge-shaped infarcts of both kidneys.

KF was continued on broad spectrum antibiotics for "culture negative" endocarditis and the question of cholycystitis. Heparin was also continued and the patient’s condition appeared to be improving until her mental status deteriorated 10/26/99. An MRI (figure 5) of her brain done 10/27/99 showed bilateral punctate infarcts in all vascular territories consistent with embolic phenomenon.

Figure 5.  MRI of brain in patient #2 (KF) with diffusion 
weighted image showing multiple, bilaterial punctate infarcts (bright lesions).

Prevalence - The post-mortem prevalence of NBTE in the general population has been examined in many autopsy series with a cited prevalence ranging from 0.3%⁹ to 9.3%¹⁰⁰. In a review of 14 autopsy series and over 80,000 patients, Lopez at al.¹¹ found the average frequency to be 1.3% in all autopsies. Kuramoto et al.¹⁰ have done the most recent, large postmortem analysis and 217 out of his 2340 (9.3%) cases had vegetations consistent with NBTE. They suggested that previous investigators underestimated the prevalence secondary to the friable nature of these lesions and inadequate handling of specimens.

It would be much more clinically useful to know the prevalence of NBTE in living patients and the widespread use of echocardiography may give us some insight into this question. Edoute et al.¹² did an echocardiographic study to try and assess the frequency and characteristics of NBTE in cancer patients. They looked prospectively at 200 non-selected ambulatory patients with solid tumors with screening transthoracic echocardiography (TTE). These patients were hospitalized for either the diagnosis of their malignancy or for chemotherapy. Completely disabled patients were excluded, as were patients with prior history of valvular disease, including rheumatic heart disease. The average time from diagnosis to TTE was 2 months. A cohort of 100 patients without known cancer or heart disease referred for evaluation of occult systemic emboli served as the control group. The researchers were blinded to the patient’s clinical history and they found that 38 of 200 cancer patients (19%) and 2 of 100 (2%) of the control group had vegetations (p<0.001).

No comparison to the gold standard (pathologic specimen) was done as this feat would be impossible in this type of study. Therefore, the true sensitivity and specificity of the echocardiographic evaluation could not be assessed. The clinical importance of these lesions is unknown as well. The authors did make an attempt to make this correlation by interviewing the patients at enrollment for the recent history or presence of thromboembolic disease and found that 9 of 38 patients (24%) with vegetations had evidence for thromboembolism compared with 13 of 162 (8%) without vegetations (p=0.013). While the results were suggestive, one interview at one point in time is probably not adequate to determine if the vegetations seen in these asymptomatic patients are relevant clinically, especially given how early in the disease course the patients were.

This study does raise questions, though, concerning the "true" prevalence of NBTE, particularly in cancer patients. As described above, the evaluation was done in the very early phase of these patients malignancy rather than the terminal phase or post-mortem phase. The prevalence may vary greatly on when in the time course of the disease it is examined. The lesions in NBTE are quite friable and are loosely attached to the valves. Perhaps as embolization occurs over time in these patients, the prevalence of vegetations would decrease. This important question remains unanswered for now. It is probably fair to say, however, that NBTE is relatively common in patients with cancer at some point along their disease course.

Age and Sex – It appears that NBTE most commonly affects patients from their fourth to eighth decades¹¹ , but cases have been reported in neonates^13,14,15, infants¹⁶, children¹⁷, and the elderly. Distribution between males and females appears to be equal.

Underlying Disease Associations – It is well established that NBTE is most commonly found in patients with malignancy. It has been found in autopsy studies to be linked with all cancers except brain tumors. Lopez et al.¹¹pooled data from all of the major autopsy series and found lung cancer to be the most common underlying malignancy, followed by pancreatic and stomach carcinoma. This is consistent with earlier reports of its close association with mucinous adenocarcinomas. It also appears to be relatively common in patients with lymphoproliferative disorders¹² and bone marrow transplant recipients^18-19. It has been associated with over 30 different underlying conditions²⁰ including most chronic diseases (i.e. autoimmune disease, renal failure, etc.), sepsis, burns²¹ , and patients with central indwelling catheters.

Recently, attention has been focused on the association of NBTE with DIC as discussed in the next section. A link between the two certainly seems to exist, and a there may even be either a causal relationship present in at least some cases. Similarly, they may be different manifestations of the same process.

Pathogenesis

The exact pathophysiology of NBTE is not completely understood. Many theories abound, but as our insight into coagulation and, more specifically, cancer associated thromboembolic disease deepens, some universal underlying themes are being advanced. Most authors believe that the following general principles illustrated by Virchow’s Triad hold true for the formation of these thrombotic vegetations: 1) alterations in blood flow at the site of thrombosis, 2) injury to the valvular endocardium, and 3) hypercoagulability (figure 6).

The valves seem to be chiefly affected at the commisures or contact margins of the valve leaflets and on the "downstream" or low-pressure side where changes in flow are likely²². There is also in vitro evidence that valvular tissue has an increased affinity for platelet binding, most likely via a prostacyclin effect²³.

Most often, the valve surfaces appear normal macroscopically, although cases of NBTE have certainly been cited in patients with underlying rheumatic heart disease, on prosthetic heart valves, and in arteriosclerotic disease¹¹ . On a microscopic level there are morphologic abnormalities of the valvular collagen indicative of edema and/or degeneration^5,7 . Proposed mechanisms for this damage have varied from "allergic damage"⁵ , vitamin deficiences²⁴, hemodynamic trauma, various noncardiac stresses²⁵, and injuries caused directly by certain neoplastic cells²⁶. Direct trauma to the valves may also lead to NBTE as evidenced by several experimental models^27-28 and in many case series of patients with centralized catheters^29-31. All of these series were relatively small, which probably contributed to the broad range of frequencies reported (3.4% to 66.7%). In all of the above circumstances, it is felt that exposed collagen on the valve surface provides a nidus for platelet adhesion and subsequent thrombus formation.

Numerous potential mechanisms for the hypercoagulable state associated with NBTE have been proposed and examined. Cancer-associated hypercoagulability has been investigated most thoroughly and both increases in procoagulant activity and decreases in native anticoagulant processes have been implicated in in the shift of the hemstatic balance toward thrombosis (figure 7). The latest insights focus on membrane bound tissue factor, which binds factor VII or VIIa and generates thrombin^32-35. Tissue factor is expressed by many malignant tissues³⁶, by mononuclear cells in response to inflammatory cytokines³⁷, and even by valvular tissue in animal models³⁸. It is known to be a triggering factor for DIC, and likely is important in the generation of NBTE (figure 8).

While controversial, it is now generally accepted that NBTE represents a manifestation of Trousseau’s syndrome, as first suggested by Sack et al in 1977³⁹. He and others have described the potential pathogenic role of a chronic thrombotic coagulopathy, namely DIC, in Trousseau’s syndrome. The relationship between DIC and NBTE has been gaining attention since 1973 when Guinn et al⁴⁰ noted the association in their small autopsy series. Many authors have since tried to better elucidate this interaction and have established a solid link between the two^10,23,40-46 . Lopez et al¹¹ summarized the data presented in autopsy series from 1973 to 1984. The prevalence of DIC in patients with pathologically proven NBTE ranged widely from 10.1% to 71%. Conversely, the prevalence of NBTE in those same patients with pathologically proven DIC ranged from 9.1% to 26% All of these studies were postmortem analyses, however, and the temporal relationships between the development of each condition is unknown. It is also difficult to compare all of these studies because of the different definitions used to classify DIC. Some required strict pathologic findings, others strict laboratory evidence, and still others both. The role of DIC as a causal mechanism for NBTE can be neither defined nor discounted based on these observations. This finding could also help explain why NBTE is seen with such a wide variety of underlying conditions.

The only attempt to examine this relationship in living patients has been undertaken by Edoute¹² . In a subset of the cancer patients mentioned previously who underwent screening echocardiography, they measured fibrin degradation products (D-dimer) by ELISA. They report that 22 of the 32 (69%) patients with vegetations that underwent D-dimer testing had a positive test, while 73 of 138 (53%) patients without vegetations had a positive test (non significant association). This study was not designed to define this relationship, and several shortcomings were evident. The determination of the presence of vegetations and the D-dimer assay were made at one point in time, both very early in the disease course of these patients. It is often difficult to make the diagnosis of chronic DIC in cancer patients as laboratory values, including D-dimers, can often fluctuate between normal and abnormal as the coagulopathy waxes and wanes. Because of these limitations, it is presently only certain that chronic DIC is often found with NBTE, but it is quite likely the causative factor in at least some cases of NBTE. In fact, many experts feel that DIC and NBTE actually represent different points in a single spectrum of disease^39,47 . Until the antemortem diagnosis of NBTE is recognized more routinely so that patients can be followed over time, it will be difficult to further our understanding of these matters.

Pathological Findings

The vegetations of NBTE are seen most often on the atrial surface of the mitral valve or the ventricular surface of the aortic valves, although all four valves can be affected¹¹ . There is an increased incidence of right sided lesions in neonates^13-16,48-49 and patients with central venous and pulmonary artery catheters^21,29-30,63 .

The lesions of NBTE are usually described as platelet and fibrin thrombi with an absence of inflammation^1,5 (figure 9). They are usually very superficial and friable. In the large autopsy series, most lesions are quite small (<4mm) and multiverrucal¹¹ (figure 10). Variants have also been reported in which larger (5-10mm) lesions are present.

Figure 10.  Variations in the valvular vegetations in NBTE.  Most are small, friable and superficially attached (left), but some are larger (right) and multi-veruccal (both).

Again it is emphasized that most of the data accrued regarding the frequency of clinical events in patients with NBTE are based on autopsy findings and only truly represent the incidence in dying patients. Systemic emboli are the main cause of morbidity in patients with NBTE and are the most clinically evident events. As summarized by Lopez et al¹¹ , systemic emboli are present in about half of the patients with NBTE at autopsy. The most commonly affected area are the brain, heart, kidneys, and mesenteric organs, yet any organ can be involved.

Neurologic events appear to be the most recognized clinical occurrences in living patients with NBTE. Deficits may range from focal losses to a more generalized confusion or "mental status change"^10,43,50-52 . Coronary embolism is also one of the complications in these patients, although it is often asymptomatic¹² . Multiple or recurrent emboli are frequent and should alert the clinician to the possibility of NBTE. Finally, concurrent venous thrombosis and pulmonary embolisms are commonly seen, especially in patients with malignancy^22,39 .

There are no pathognomonic physical findings for NBTE. Cardiac murmurs may be noted, but are only evident about one third of the time in most series^50,53-55 . Laboratory evidence for NBTE is usually in the form of a coagulation disturbance, most often DIC. Some authors have noted that when DIC and NBTE are documented together, the incidence of systemic emboli is much higher^12,40 . Similarly, it has been documented that in patients with known neoplasia and associated chronic DIC with systemic emboli, that NBTE is present at autopsy in as many as 74% of patients³⁹.

It has also been proposed that NBTE may be important as a substrate or initiating event for infective endocarditis⁵⁶ . This argument stems from animal models where bacterial endocarditis can be produced in rabbits with NBTE by a single injection of bacteria^27,57 . This relationship is not well understood in humans as of yet.

The cases presented earlier serve as excellent examples of the salient clinical features that can be seen with NBTE. Both had neurologic manifestations as their presenting symptom. Both also had multiple systemic infarcts, namely renal, splenic, and likely gastric. RR had no cardiac murmur while KF had a new mitral regurgitation murmur. A D-dimer assay was not available for RR, but good evidence for DIC existed for KF. Unfortunately both of these patients were in the terminal stages of their respective diseases and ultimately died from complications of NBTE, but this serves to remind us that it is a clinically important entity that needs to be recognized and treated.

Diagnosis

NBTE has been discussed at clinicopathologic conferences for many years, but has rarely been considered at the bedside of patients. It is now clear that it is not a rare, unimportant problem without clinical relevance, especially in certain patient populations. To make the diagnosis in living patients so that rational therapy can be implemented, clinicians must think of and seek out NBTE more often. With our increasing understanding of the pathophysiology of NBTE and the advent of higher quality and more available echocardiography, perhaps the diagnosis can be made in time to avert some of the devastating clinical complications just discussed.

The following three criteria were proposed for making the clinical diagnosis of NBTE in the 1950’s⁵⁸ : 1) the presence of an underlying condition known to be associated with NBTE, 2) the presence of heart murmurs, and 3) clinical evidence of multiple arterial embolizations involving the brain, spleen, or kidneys. With today’s more advanced imaging studies and greater insight into thrombotic phenomenon, these criteria are perhaps antiquated. We now know that the presence or absence of heart murmurs is not predictive for NBTE. The rise of echocardiography may also allow for the diagnosis of NBTE prior to systemic embolization. The interconnectedness of NBTE, DIC and Trousseau’s syndrome may allow for identification of high risk patients who can be examined more closely. Finally, the early diagnosis of NBTE in patients without known malignancy or chronic disease may allow for a search of a potentially curable occult process.

Echocardiography – In 1976, Estevez and Corva⁵⁹ presented a case report of a patient with NBTE that was detected with M-mode echocardiography who subsequently underwent successful valve replacement. Since that time several investigators have reported the diagnosis by 2-D echo⁶⁰ . The only group that has made an attempt to study the usefulness of echo for detecting NBTE in living patients is Edout et al.¹² As discussed earlier, they performed screening echocardiograms of 200 hospitalized cancer patients and found 38 (19%) of them to have vegetations. The limitations of this study as mentioned earlier was chiefly the single evaluation so early in the disease course. We are not able to draw conclusions from this study about which of these vegetations caused morbidity in these patients. No study has been done that follows asymptomatic patients with NBTE prospectively over time to note their clinical course. Likewise, no study has been done to determine the performance characteristics of echocardiography for diagnosing NBTE.

More recently, Blandchard et al.⁶¹ report a case of NBTE diagnosed by transesophageal echocardiography (TEE), as was the case in our two case presentations earlier. The small nature of many of these lesions may not be seen as readily by TTE in certain patients secondary to poor positioning or suboptimal quality or resolution. TEE is generally felt to provide superior images and greater detail of cardiac chambers and valves⁶² .

It seems to be clear that echo can help confirm the diagnosis of NBTE when it is expected by clinicians. It may allow for diagnosis early enough to help patients. TEE should be considered if other test are non-diagnostic and the clinical suspicion remains high.

Laboratory Evidence and Other Modalities – The relationship between NBTE and DIC was discussed previously. The presence or absence of laboratory evidence of DIC cannot "rule-in" or "rule-out" NBTE, but it can produce another large piece of the clinical puzzle. Laboratory evidence targeted at symptoms can also be helpful, i.e., hematuria for renal infarcts or abnormal liver function test for hepatic infarcts. Also, infective endocarditis always needs to be excluded, so serial blood cultures should be obtained.

While the roles of computed tomography and magnetic resonance imaging are well-established in documenting cerebral and other systemic infarcts, their utility in evaluating valvular vegetations is not known. Other techniques such as radio-labeled platelets are currently undergoing investigation for evaluating the lesions of infective endocarditis⁶³ and, if useful, may be helpful in NBTE.

Therapy

It is not surprising that no prospective controlled trials have been done to assess treatment options for NBTE given the rarity of antemortem diagnosis. Now that we are able to recognize this problem in living patients, perhaps this will give us an opportunity to investigate this avenue. Obviously, diagnosis must precede any rational therapy. Once NBTE is recognized, therapeutic approaches could center on either the underlying cause of NBTE (i.e. malignancy, sepsis, etc.), or the thromboembolism and its associated hypercoagulable state. This discussion will focus on the latter.

Potential therapeutic options for currently treating the thromboembolic process and underlying hypercoagulability include anti-platelet agents, thrombolytic therapy, coumadin, heparin, or low-molecular-weight heparin (LMWH). There have also been at least two successful surgical interventions reported in the literature^59,64. Until a study is done looking at patients with NBTE and their response to therapy over time, clinicians are left with inference, scientific theory, anecdotal evidence, and art to care for them. As discussed earlier, NBTE, DIC, and Trousseau’s syndrome are considered by some to the be a spectrum of the same process at different time points. With this concept in mind, present therapies for both DIC and Trousseau’s will be discussed.

DIC – The hallmark of therapy for DIC is treatment of the underlying disorder, as it is currently not clear that it can exists as a primary process. Measures geared at treating the DIC itself may be necessary, but supportive evidence is scarce. Heparin has been shown to have a beneficial effect in some small, uncontrolled studies, but has failed to show this benefit in larger, controlled trials^65,66. Heparin does not appear to significantly increase bleeding risk in these patients. These investigations focused on acute DIC, but a more chronic or compensated form is presumably involved with NBTE. Anecdotally, experts describe more success in treating this chronic form of DIC, where thrombosis predominates without the hemorrhagic complications secondary to consumption of clotting factors. This observation makes intuitive sense pharmacologically. There are investigators that believe that LMWH may be used as an alternative in DIC⁶⁷.

Replacement of clotting factors, platelets, and physiologic anti-coagulants appears promising in patients with acute DIC, but their efficacy for a more chronic process as seen in NBTE is unknown. The cost of these products will probably preclude them from being used on a long-term basis for now. Future options such as newer inhibitors of thrombin like desrudin, tissue-factor inhibitors like recombinant nematode anticoagulant protein c2, and activated protein C supplementation are currently being investigated⁶⁸.

Trousseau’s Syndrome - Patients with malignancies often have isolated thromboses following surgery or during the administration of certain chemotherapeutic agents⁶⁹. These events usually respond to traditional therapy with heparin or LMWH initially, followed by longer term coumadin. Less common is another subset of cancer patients who have a more aggressive thrombotic diathesis termed Trousseau’s syndrome. The generally accepted features of the syndrome include spontaneous recurrent or migratory episodes of venous thrombosis and arterial emboli in the setting of a malignancy. Although controversial, NBTE is thought to be a variant of the syndrome.

In 1977 Sack et al³⁹ reported an extensive summary of the clinical, pathophysiologic, and therapeutic features of Trousseau’s syndrome. He retrospectively reported the findings of 55 cancer patients with recurrent thrombosis or arterial emboli that received anticoagulants. Enthusiasm for his results must be tempered given the small retrospective analysis, unknown patient characteristics, lack of a control group, uncertainty of temporal relations of events, and loosely defined criteria for response or failure. Nonetheless, he seemed to show a trend that heparin was much more successful than coumadin in this patient population. In fact, 20 of 24 patients given heparin alone responded (defined as "cessation of thrombophlebitis, arterial emboli or hemorrhage") while only 2 of 10 responded to coumadin alone. A group of 22 patients were given both heparin and coumadin at different times and 12 responded to heparin and failed coumadin, 4 responded to both, and 6 failed both. He also noted that 55% of patients who responded to heparin had recurrent events upon cessation of therapy, and that no one who failed heparin responded to coumadin. He and numerous other investigators since then have repeatedly documented anecdotally that coumadin failed to prevent embolic events in this group of patients, while heparin was usually effective^39,70-72. The reason for this in unknown. More recently, other investigators have demonstrated that LMWH may be effective and present another option for long-term management in this subset of patients^73-74. One group describes successfully treating these patients for as long as 27 months without adverse events until the patients were lost to follow-up.

The safety of heparin or LMWH administration in patients with Trousseau’s syndrome is not well established. One group of investigators looked retrospectively at patients with NBTE and cerebral infarction treated with heparin and found no increased risk for bleeding, but this was in a small set of patients⁷⁵. In patients with cerebral infarctions from other types of cardiac disease, the risk of intracerebral hemorrhage appears less than the risk of re-embolization, except in patients with very large infarctions or hemorrhagic infarctions⁷⁵. Other complications of heparin therapy including heparin-induced thrombocytopenia, osteoporosis, and hypoaldosteronism have not been examined in this set of patients. Likewise, the safety profile of LMWH in patients with NBTE is not known but it appeared to be at least as safe as heparin in several recent large trials where it was used in thromboembolic disease, DVT prophylaxis, and unstable angina⁷³. It also appeared in these trials that fewer patients developed heparin-induced thrombocytopenia with LMWH. Since patients with NBTE would likely need long-term therapy, LMWH should be considered given its ease of administration and apparent efficacy and safety.

Summary and Key Points

The epidemiology, pathogenesis, clinicopathologic features, diagnostic modalities, and treatment options of NBTE have been discussed. Two cases diagnosed antemortem were presented to underscore important clinical features of the process. The reader should take away several points including the following:

1. Think of NBTE. Diagnosis and therapy cannot be initiated unless clinicians are aware of NBTE and seek it out. All cases of "culture-negative endocarditis" should be examined closely. Patients with spontaneous thromboembolic disease not otherwise explained, with recurrent thromboses or emboli, with DIC and malignancy, and with thromboses and malignancy, should all be considered higher risk populations.
2. Diagnose NBTE. Clinicians should recognize high risk patients and look for supportive laboratory evidence for NBTE. Echocardiography can be used to confirm clinical suspicion. Transesophageal approaches may be more helpful in some patients.
3. Treat NBTE. While there is a paucity of data to help clinicians decide on therapeutic options, heparin and LMWH seem to offer a viable option for preventing some of the devastating complications of NBTE. In patients found to have NBTE without a known underlying disease, a search for a treatable occult process should be undertaken.

Many unanswered questions still abound. Should asymptomatic patients with adenocarcinomas undergo routine echocardiography? Should high risk patients (as described above)? Can laboratory parameters, i.e. D-dimers, be used as a predictor for clinical events or to guide therapy? What is the best agent and optimal length of therapy? Should asymptomatic patients with vegetations be treated? Can or should patients with terminal processes be treated to prevent symptoms? All of these questions are important and deserve exploration. In order for them to be answered, clinicians need to identify patients with NBTE so that they may be followed clinically, hopefully in prospective clinical trials. Until then, care plans must be individually tailored on a case by case basis.

1. Gross L, Friedberg CK. Nonbacterial thrombotic endocarditis. Classification and general description. Arch Intern Med 1936:58:620-640.
2. Zeigler R. Uber den bau die entstehung endocaritischer efflorescenzen. Werh Dtsch Kong. Intern Med 1888;7:399.
3. Harbitz F. Studen uber endocarditis. Dtsh Med Wochenschr 1899;25:121.
4. Libman E. Characterization of various forms of endocarditis. JAMA 1923;80:813-818.
5. Allen AC, Sirota JH. The morphogenesis and significance of degenerative verrucal endocarditis. Am J Pathol 1944;201025-1055.
6. Kossman CE. Nonbacterial thrombotic endocarditis: A specific lesion of Trousseau’s Syndrome. J of the Tennessee Med Assoc 1980;73(1):14-16.
7. Angrist A, Marquiss J. The changing morphologic picture of endocarditis since the advent of chemotherapy and antibiotic agents. Am J Pathol 1954;30:39-63.
8. MacDonald RA, Robbins SL. The significance of nonbacterial thrombotic endocarditis: an autopsy and clinical study of 78 cases. Ann Intern Med 1957;46:255-273.
9. Rohner RF, Prior JT, Sipple JH. Mucinous malignancies, venous thrombosis and terminal endocarditis with emboli: a syndrome. Cancer 1966;19:1805-1812.
10. Kuramoto K, Matsushita S, Yamanouchi H. Nonbacterial thrombotic endocarditis as a cause of cerebral and myocardial infarction. Jpn Circ J 1984;48:1000-1006.
11. Lopez JA, Ross RS, Fishbein MC. Nonbacterial thrombotic endocarditis: a review. Am Heart J 1987;113:773-784.
12. Edoute Y, Haim N, Brenner B. Cardiac valvular vegetations in cancer patients: A prospective echocardiographic study of 200 patients. Am J of Med 1997;102:252-258.
13. McDonald RH. Valvular thrombotic vegetation in newborn. Arch Pathol 1950;50:538-544.
14. Krous HF. Neonatal nonbacterial thrombotic endocarditis. Arch Pathol Lab Med 1979;103:76-78.
15. Morrow WR, Hass JE, Benjamin DR. Nonbacterial endocardial thrombosis in neonates: relationship to persistent fetal circulation. J Pediatr 1982;100:117-122.
16. Liwnicz BH, Lepow H. Nonbacterial thrombotic endocarditis in premature child. Clinical significance and possible relationships to subvalvular hematoma. NY Stat J Med 1976;76:912-916.
17. Young RSK, Zalneraitis EL. Marantic endocarditis in children and young adults: clinical and pathologic findings. Stroke 1981;12:635-639.
18. Jerman MR, Fick RB. Nonbacterial thrombotic endocarditis associated with bone marrow transplantation. Chest 1986;90:919-922.
19. Patchell RA, Whit CL, Clark AW. Nonbacterial thrombotic endocarditis in bone marrow transplant patients. Cancer 1985;55:631-635.
20. Biller J. Nonbacterial thrombotic endocarditis, Arch Neurol 1982;39:95-98.
21. Ehrie M, Morgan AP, Moore FD, O’conner NE. Endocarditis with indwelling balloon-tipped pulmonary artery catheters in burn patients, J Trauma 1978;18:664-666.
22. Deppisch LM, Fayemi AO. Nonbacterial thrombotic endocarditis. Clinicopathologic correlations. Am Heart J 1976;92:723-729.
23. Johnson CM, Helgeson SC. Platelet adherence to cardiac and noncardiac endothelial cells in culture: Lack of a prostacyclin effect. J of Lab and Clinical Med 1988;112(3):372-379.
24. Rinehart JF, Mettier SR. Heart valves and muscles in experimental scurvy with superimposed infection with notes on similarity of lesions to those of rheumatic fever. Am J Pathol 1964;19:135-143.
25. Angrist AA, Oka M, Nakao K, Marquiss J. Studies in experimental endocarditis. I. Production of valvular lesions by mechanisms not involving infection or sensitivity factors. Am J pathol 1960;36:181-199.
26. Warren B, Vales O. The adhesion of thromboplastic tumor emboli to vessel wall in vitro. Br J Exp. Pathol 1972;153:301-313.
27. Durack DT, Beeson PB. Experimental bacterial endocarditis. I. Colonization of a sterile vegetation. Br J Exp Pathol 1973;54:142-151.
28. Gutschik E, Christenson N. Experimental endocarditis in rabbits I. Technique and spontaneous course of nonbacterial thrombotic endocarditis. Acta Pathol Microbiol Scand 1978;86:215-221.
29. Green JF, Cummings KC. Aseptic thrombotic endocardial vegetations. A complication of indwelling pulmonary artery catheters. JAMA 1973;225:1525-1526.
30. Pace NL, Horton W. Indwelling pulmonary artery catheters. Their relationship to aseptic thrombotic endocardial vegetations. JAMA 1975;233:893-894.
31. Rowley KM, Clubb KS, Smith GSW, Cabin HS. Right sided infective endocarditis as a consequence of flow directed pulmonary artery catheterization. A clinicopathological study of 55 autopsied patients. N Engl J Med 1984;311:1152-1156.
32. Kakkar AK, DeRuvo N, Chinswangwatanakul V, Tebbult J, Williamson RCN. Extrinsic factor activation in cancer with high factor VIIa and tissue factor. Lancet 1995;346:1004-1005.
33. Lindahl AK, Abildgaard U, Stokke F. Release of extrinsic pathway inhibitor after heparin injection: increased response in cancer patients. Thromb Res 1990;59:651-656.
34. Moake JL. Hypercoagulable states. Adv Inter Med 1990;35:235-248.
35. Edwards RL, Rickles FR, Cronlund M. Abnormalities of blood coagulation in patients with cancer: mononuclear cell tissue factor generation. J Lab Clin Med 1987;98:917-928.
36. Callander N, Rappaport S. Trousseau’s syndrome. The Western J of Med 1993;158:364-371.
37. Osterud B, Flaegstad T. Increased tissue thromboplastin activity in monocytes of patients with meningogoccal infection: related to an unfavorable prognosis. Thromb Haemost 1983;49:5-7.
38. Nakanishi K, Fumiko T, Nakata Y, et al. Tissue factor is associated with the nonbacterial thrombotic endocarditis induced by a hypobaric hypoxic environment in rats. Virchows Archive 1998;433(4):375-379.
39. Sack GH, Levin J, Bell WR. Trousseau’s syndrome and other manifestations of chronic disseminated coagulopathy in patients with neoplasms: clinical, pathophysiologic and therapeutic features. Medicine 1977;56:1-37.
40. Guinn GA, Ayala A, Liddicoat J. Clinical and therapeutic considerations in nonbacterial thrombotic endocarditis. Chest 1973;64:26-28.
41. Kim H-S, Suzuki M, Lie JT, Titus JL. Nonbacterial thrombotic endocarditis and disseminated intravascular coagulation. Autopsy study of 36 patients. Arch Path Lab Med 1977;101:65-68.
42. Bedikan A, Valdiviesco M, Luna M, Bodey GP. Nonbacterial thrombotic endocarditis in cancer patients: comparison of patients with and without concomitant disseminated intravascular coagulation. Med Pediatr Oncol 1978;4:149-157.
43. Biller J, Chall VR, Toole JF, Howard VJ. Nonbacterial endocarditis. A neurologic perspective of clinicopathologic correlations of 99 patients. Arch Neurol 1982;39:95-98.
44. Rabboy SJ, Colman RW, Minna JD. Pathology of disseminated intravascular coagulation. Analysis of 26 cases. Hum Pathol 1972;3:327-343.
45. Goodnight SH. Bleeding and intravascular clotting in malignancy: a review. Am NY Acd SCI 1974;230:271-288.
46. Sugiura M, Hiraoka K, Ohkawa S-I, Ueda K, Matsuda T, Murakami M. A clinicopathologial study on cardiac lesionsin 64 cases of disseminated intravascular coagulation. Jpn Heart J 1977;18:57-69.
47. Kossmann CE. Nonbacterial thrombotic endocarditis: A specific lesion of Trousseau’s syndrome. J of Tenn Med Assoc 1980;73(1):14-16.
48. Kunstadter RH, Kaltenekker F. Acute verrucous endocarditis in the newborn. J Pediatr 1962;61:58-64.
49. Favara BE, Franciosi RA, Butterfiled LJ. Disseminated intravascular and cardiac thrombosis of the neonate. Am J Dis Child 1974;127:197-204.
50. Kooiker JC, Maclean JM, Sumi SM. Cerebral embolism, marantic endocarditis and cancer. Arch Neurol 1976;33:260-264.
51. Mackenzie TB, Popkin MK. Psychological manifestations of nonbacterial thrombotic endocarditis. Am J Psychiatry 1980;137:972-973.
52. Remillard GM, Carpenter S. Cerebral embolism due to nonbacterial thrombotic endocarditis. Can Med Assoc J 1972;106:573-576.
53. Rosen P, Armstrong D. Nonbacterial thrombotic endocarditis in patients with malignant neoplastic disease. Am J Med 1973;54:23-26.
54. Bryan CS. Nonbacterial thrombotic endocarditis with malignant tumors. Am J Med 1969;46:1000-1006.
55. Waller BF, Knapp WS, Edwards JD. Marantic valvular vegetations. Circulation 1973;48:644-650.
56. Angrist AA, Oka M. Pathogenesis of bacterial endocarditis. JAMA 1963;183:249-120.
57. Durack DT, Beeson PM, Petersdorf RG. Experimental bacterial endocarditis. III. Production and progress of the disease in rabbits. Br J Exp Pathol 1973;54:142-151.
58. McKay DG, Wahler GH. Disseminated thrombosis in colon cancer. Cancer 1955;8:970-978.
59. Estevez CM, Corva BC. Serial echocardiographic abnormalities in nonbacterial thrombotic endocarditis of the mitral valve. Chest 1976;69:801-804.
60. Lopez JA, Fisbein MC, Siegel RJ. Echocardiographic features of nonbacterial thrombotic endocarditis. Am J Caardiol 1987;59:478-480.
61. Blanchard DG, Ross RS, Dittrich HC. Nonbacterial thrombotic endocarditis. Assessment by transesophageal echocardiography. Chest 1992;102(3):954-956.
62. Mugge A, Daniel WG, Frank G, Lichtlen PR. Echocardiographic diagnosis of infective endocarditis: reassessment of prognostic implications of vegetation size determined by the transthoracic and transesophageal approach. J Am Coll Cardiol 1989;14:631-638.
63. Ribe AL, Thakur ML, Gottschalk Aa, Andriole VT, Zaret BL. Imaging experimental infective endocarditis with indium 111-labeled blood cellular components. Circulation 1979;59:336-343.
64. Habbab M, Al-Zaibag MA, Al-Hilali AM, Al-Fagih MR. Unusual presentation and echocardiographic features of surgically proven nonbacterial thrombotic endocarditis. Am Heart J 1990;119:404-406.
65. Corrigan JJ. Heparin therapy in bacterial septicemia. J Pediatr 1977;91:695-700.
66. Feinstein DI. Diagnosis and management of DIC: the role of heparin therapy. Blood 1982;60:284-287.
67. Sakuragawa N, Hasegawa H, Maki M, Nakagawa M, Nakashima M. Clinical evaluation of LMWH on DIC – a multicenter co-operative double-blind trial in comparison with heparin. Thromb Res 1993;72:475-500.
68. Levi M, Cate H. Disseminated intravascular coagulation. New England J Med 1999;341:586-592.
69. Levine MN, Gent M, Hirsh J, Goodhyear MD, Hrtniuk W, et al. The thrombotic effect of anticancer drug therapy in women with stage II breast cancer. N Engl J Med 1988;318:404-408.
70. Mosesson MW, Coleman RW, Sherry S. Chronic intravascular syndrome. N Engl J Med 1968;278:815-821.
71. Bell WR, Starksen NF, Tong S, Porterfiled JK. Troussaeu’s syndrome – Devastating coagulopathy in the absence of heparin. Am J Med 1985;79:423-430.
72. Kazmier FJ, Bowie EJW, Hagedorn AB, Owen CA. Treatment of intravascular coagulation and fibrnolusis syndrome. Mayo Clin Proc 1974;49:665-672.
73. Walsh-McMonagle D, Green D. Low-molecular-wight heparin in the management of Trousseau’s syndrome. Cancer 1997;80:649-655.
74. Zuger M, Demarmels Biasiutti F, Wuillemin WA, Furlan M, Lammle B. Subcutaneous LMWH for treatment of Trousseau’s syndrome. Annals of Hematology 1997;75(4):165-167.
75. Rogers LR, Cho E, Kempin S, Posner JB. Cerebral infarction from nonbacterial thrombotic endocarditis. Clinical and pathological Study Including the effects of anticoagulation. Am J Med 1987;83:746-756