Risk factors for “major” embolic events in hospitalized patients with infective endocarditis

Risk factors for “major” embolic events in hospitalized patients with infective endocarditis Emanuele Durante Mangoni, MD,a Luigi E. Adinolfi, MD,a Marie-Francoise Tripodi, MD,a,b Augusto Andreana, MD,a Michele Gambardella, MD,a Enrico Ragone, MD,a Davide F. Precone, MD,a Riccardo Utili, MD,a,b and Giuseppe Ruggiero, MDa Naples, Italy

Background Infective endocarditis often is complicated by embolic events after hospital admission. Identifying patients at higher risk may improve the disease outcome. This study was aimed at identifying predictors of embolic risk among the clinical and laboratory data obtained on hospital admission in patients diagnosed as having definite infective endocarditis according to the Duke criteria. Methods

Ninety-four patients were enrolled in a prospective study. The results of hematologic, echocardiographic, and microbiological investigations were analyzed, using statistical methods as appropriate. Multivariate analysis was applied to variables significantly associated with embolism in univariate analysis.

Results Forty-six percent of patients had a major embolic complication after admission. No association was found between embolism and sex, site of infection, or microorganism involved. Patients with embolism were significantly younger, had larger vegetation, and showed a significantly higher level of serum C-reactive protein and lower albumin concentrations than those without embolism. Young age, larger vegetation size, and high levels of C-reactive protein were the independent variables associated with an increased incidence of embolic events in the multivariate logistic regression analysis. Conclusions Our data indicate that patients with infective endocarditis with young age and/or with large vegetation and/or with high serum levels of C-reactive protein are at increased risk of major embolic complications during the in-hospital course of the disease. (Am Heart J 2003;146:311– 6.)

See related Editorial on page 189.

Infective endocarditis (IE) is a life-threatening clinical condition that is still responsible for a substantial mortality rate despite the great progress made in the diagnosis and treatment.1,2 Systemic embolism is among the primary causes of morbidity and mortality in this setting. Embolic events during IE are observed with a frequency approaching 50% of cases in prospective, targeted investigations.3 When causing overt clinical manifestations, these acute complications may be defined as “major” embolic events. However, it has been demonstrated that a substantial number (38%) of embolic events are not symptomatic and are often overFrom the aInstitute of Medical Therapy, and bResearch Centre for Cardiovascular Disease, Second University of Naples, Italy. Supported in part by a grant from MIUR, Italy, to the Research Centre on Cardiovascular Disease and in part by a grant from the Second University of Naples. E. Ragone is a fellow student of the PhD course of Cardiological and Cardiosurgical Sciences. Submitted November 29, 2001; accepted May 17, 2002. Reprint requests: Luigi E. Adinolfi, MD, Istituto di Terapia Medica, Seconda Universita` di Napoli, c/o Osp. Gesu` e Maria, Via Cotugno 1, 80135, Napoli, Italy. E-mail: [email protected]. © 2003, Mosby, Inc. All rights reserved. 0002-8703/2003/$30.00 ⫹ 0 doi:10.1016/S0002-8703(02)00093-5

looked in the clinical setting.4 It is reasonable to believe that such “minor” embolic events do not significantly modify the clinical course of the disease. Conversely, evidence has accumulated to suggest that “major” embolic events, such as cerebral embolism, may complicate the illness,5 worsen the prognosis,6 and affect the mortality rate.7–11 Thus, identifying patients at higher risk for such complications is of crucial importance to improve the disease outcome. Several studies aimed at identifying predicting factors of embolism in IE have been performed, but conflicting results have been reported.12–26 However, many of these studies were conducted before the introduction and validation of the Duke criteria for the diagnosis of IE.27 Thus, the contradictory results might be due to a patient selection bias. The aim of this study was to investigate prospectively the role of several clinical, biochemical, and instrumental parameters on embolic risk in patients with definite IE during the inhospital stay.

Methods From May 1995 to December 2000, all consecutive patients who were referred to our center for suspicion of IE or for a fever of unknown origin were evaluated for endocarditis.

American Heart Journal August 2003

312 Durante Mangoni et al

Ninety-four patients who met the diagnosis of definite IE according to the Duke criteria27 were prospectively included in the study. Nine (9%) additional patients with IE, observed in the same period, who showed on admission signs or symptoms or evidence of a previous embolic event, were excluded. The duration of fever before hospital admission was recorded for all patients. The following laboratory tests were performed at hospital admission on peripheral venous blood samples: full blood count, prothrombin time, erythrocyte sedimentation rate (ESR), dosage of serum proteins, fibrinogen, C-reactive protein (CRP), and serum ferritin. A full blood count was performed with a laser “light-scattering” particle counter (H2 Technicon, Bayer, Milano, Italy). The prothrombin time was evaluated as International Normalized Ratio (INR). The dosage of serum proteins was performed by electrophoresis; serum fibrinogen concentration was determined by nephelometry. Serum CRP and ferritin were assayed by immunoradiometric assays. Personnel unaware of the patient’s clinical data performed all tests at the University Hospital laboratory. Four to 6 blood cultures (BBL Septi-Chek, Becton Dickinson, Le Pont de Claix, France) were obtained for all patients before starting antibiotics and were incubated aerobically and anaerobically. Isolated pathogens were identified with standard methods (Gram stain, oxidase, and catalase tests) and biochemical reactions (API system, BioMe`rieux, Paris, France). On admission to the hospital, all patients underwent transthoracic and/or transesophageal echocardiography, chest radiography, and abdominal ultrasound scan. Two-dimensional echocardiography was applied to reveal vegetation on valve leaflets, and the vegetation size was measured as the length of the echogenic image, that is, its greatest diameter. In cases with multiple vegetation, the value of the length of the longest vegetation was considered. During the inhospital stay, we carefully evaluated patients with signs or symptoms suggestive of an embolic event, such as sudden neurologic or visual dysfunction, abdominal pain, hematuria, chest discomfort, abrupt dyspnea, or limb pain. Computed tomography scanning or nuclear magnetic resonance imaging were performed to confirm the clinical suspicion of cerebral embolism. Abdominal evaluation was done with ultrasound and/or computed tomography scans. Pulmonary embolism was evaluated by chest radiography, computed tomography, and perfusion radionuclide lung scans. Embolic events involving the vascular periphery were assessed by echocardiographic color Doppler examination. Statistical analysis was performed on data obtained on hospital admission, comparing patients with and those without embolic events. Possible risk factors for embolism included 17 variables: age, sex, culture positivity, isolated organism, vegetation size, site of infection, type of valve involved (native or prosthetic), prothrombin time, hemoglobin, white blood cell (WBC) count, platelet count, levels of serum fibrinogen, albumin, ESR, CRP, ferritin, and ␣2-globulins. However, to reduce spurious associations, due to a low number of events, the risk factors were cut down, eliminating variables that showed multiple correlation by Pearson linear correlation. Homogeneity of variance was assessed by using the Levene test. Accordingly, the difference between embolic and nonembolic groups was evaluated by heteroscedastic t test, homoscedastic t test, or ␹2 test, as appropriate. Quartiles

of the range of variables were utilized as cutoffs in the univariate analysis. Adjusted odds ratios (ORs) were calculated by multiple logistic regression analysis to identify variables that were independently associated with embolism. In the logistic model, embolism was the outcome variable and only variables significantly associated with embolism in the univariate analysis were included in the multivariate analysis. Unless otherwise specified, data are presented as mean ⫾ SD. A P value of ⱕ.05 was assumed to denote significance. The data analysis was carried out with the use of SPSS software (SPSS Inc, Chicago, Ill).

Results The demographic, clinical, and microbiological characteristics of the 94 patients included in the study are shown in Table I. The table reports 14 of the 17 risk factors of embolism initially planned, because 3 (hemoglobin, ESR, and ferritin) were excluded from the statistical analysis due to their high correlation with other variables. Seventy percent of patients had endocarditis on a native valve. Bacteria were isolated from blood cultures in 78 of the 94 cases (83%). Sixteen patients had negative blood cultures, presumably because of previous use of antibiotics, but in all it was possible to make a diagnosis of “definite” IE, on the basis of a major criteria, that is, endocardial involvement (vegetation, 8; new partial dehiscence of prosthetic valve, 6; new valvular regurgitation, 1; abscess, 1) and 3 minor criteria, that is, fever (⬎38.0°C) and predisposing heart condition in all patients; Osler nodes in 4 patients; rheumatoid factor in 5 patients; glomerulonephritis in 6 patients; and Janeway lesions in 1. Ninetytwo patients showed clear evidence of endocardial involvement on 2-dimensional echocardiography (transthoracic and/or transesophageal). Of these, 82 had vegetation. In the 2 patients without major evidence of endocardial involvement, the diagnosis of IE was made according to positive blood cultures for typical microorganism and 3 minor criteria. Aortic location (46%) and streptococcal cause (44% of culture positive cases) prevailed. The median interval between onset of fever and hospital admission was 48 days (3–300 days). Most of the patients (91%) had received empirical antibiotic treatment before hospital admission. The median hospitalization time was 28 days, with a range of 7 to 60 days. Overall, 75 symptomatic embolic events occurred in 43 of 94 patients (46%). Eighty-eight percent of the embolic events occurred within the first 3 weeks after admission. Twenty-two episodes involved the vascular periphery, 20 the central nervous system, 15 the spleen, 12 the lung, and 6 the kidney. The demographic and clinical features of the patients according to the presence or absence of embolic events are reported in Table I. Patients who had an embolic event were significantly younger than those who did not

American Heart Journal Volume 146, Number 2

Durante Mangoni et al 313

Table I. Demographic, clinical and microbiological characteristics of the study population and laboratory findings Parameters No. Median age (y) (range) Sex (male/female) Type of valve (%) Native valve Prosthetic valve Site (%) Aortic Mitral Aortic ⫹ mitral Right-sided IE Median vegetation size (mm) (range) Blood cultures (%) Positive Negative Pathogen (%) Streptococcus Staphylococcus Enterococcus Gram negative Others Laboratory findings* WBC (N ⫻ 103/␮L) Platelets (N ⫻ 103/␮L) INR Albumin (g/dL) ␣2-Globulins (g/dL) Fibrinogen (mg/dL) CRP (mg/L)

Nonembolic patients

P

43 39 (7–72) 27/16

51 52 (15–77) 32/19

.003† NS‡

66 (70) 28 (30)

35 (81) 8 (19)

31 (61) 20 (39)

43 (46) 32 (34) 10 (11) 9 (9) 10 (1.3–38)

20 (46) 14 (33) 3 (7) 6 (14) 14 (1.3–38)

23 (45) 18 (35) 7 (14) 3 (6) 7 (2–27)

78 (83) 16 (17)

38 (88) 5 (12)

40 (78) 11 (22)

NS‡

34 (44) 19 (24) 17 (22) 7 (9) 1 (1)

17 (45) 11 (29) 8 (21) 2 (5) –

17 (43) 8 (20) 9 (22) 5 (13) 1 (2)

NS‡

All patients 94 46.5 (7–77) 59/35

9.2 (1.75–25.7) 242 (40–758) 1.14 (0.9–3.7) 3.4 (1.4–5.2) 1 (0.48–1.77) 423 (107–1000) 29 (3.5–429)

Embolic patients

10.6 (3.77–25.7) 242 (42–758) 1.12 (0.9–1.63) 3.1 (1.4–4.5) 1 (0.48–1.77) 455 (240–838) 46 (3.5–429)

8.8 (1.75–20.5) 243 (40–635) 1.15 (0.9–3.7) 3.6 (2.3–5.2) 0.96 (0.5–1.42) 401.5 (107–1000) 18 (5–137)

NS‡

NS‡ .005†

NS† NS† .006§ .001† NS† NS† .007†

*Values represent median (range). †Homoscedastic t test. ‡␹2 Test. §Heteroscedastic t test.

(median age, 39 and 52 years, respectively, P⬍.003) and showed a higher serum CRP value (62.2 ⫾ 76.8 and 29.6 ⫾ 31.4 mg/L, respectively, P⬍.02). The mean serum albumin level was significantly lower in subjects with embolism than in patients without embolism (3.2 ⫾ 0.6 and 3.6 ⫾ 0.6 g/dL, respectively, P⬍.001). Patients without embolic events showed a lower prothrombin activity than those with embolic events (INR, 1.16 ⫾ 0.15 vs 1.51 ⫾ 0.7, P⬍.006). However, we found a higher prevalence of patients with a prosthetic valve who were receiving warfarin treatment in the group without embolic events than in those with (39% vs 19%, respectively, P ⫽ .051). There was no difference in the incidence of embolic events between patients with multiple vegetations and those with single vegetation (28% vs 52%, respectively, data not shown). Mean vegetation size was greater in patients with embolic events than in those without (15 ⫾ 9.2 and 10.2 ⫾ 7.1 mm, respectively, P⬍.03). There was no difference in the mean interval between the onset of fever and hospital admission between patients who had a major embolic event and those without embo-

lism (45 days and 52.5 days, respectively, P ⫽ NS). No individual etiologic agent was associated with a higher incidence of embolic events. Table II shows the analysis of risk for individual factors potentially associated with embolism. The cutoff to evaluate the RR was quartiles of range. The RR of having a major embolic event for patients with age under the third quartile (47–58 years) was significantly higher that of those over the 3rd quartile (Table II, P ⫽ .017). Patients with a high CRP level (3rd and 4th quartiles) were at increased risk for embolism (Table II, P ⫽ .043). In addition, values of albumin under the 3rd quartile (3.5–3.7 g/dL) were associated with embolic events with a RR higher than values of the 4th quartile (Table II, P ⫽ .005). Finally, a significantly higher risk of embolism was associated with the increased vegetation size over the 1st quartile (Table II, P ⫽ .041). No other categorical data were associated with an increased risk for a major embolic event. Table III shows the results of the logistic regression analysis performed by using the 4 variables associated

American Heart Journal August 2003

314 Durante Mangoni et al

Table II. Univariate analysis of potential risk factors for embolism Variables

Quartiles

Percent embolic patients

Relative risk (95% CI)

Age (y)

⬍29 29–46 47–58 ⬎58 ⬍6.5 6.5–10 10.5–15 ⬎15 ⬍1.1 1.1–1.14 1.15–1.27 ⬎1.27 ⬍2.9 2.9–3.4 3.5–3.7 ⬎3.7 ⬍10.5 10.5–29 29.1–61 ⬎61

66.7 56.5 29.2 30.4 18.8 52.2 61.5 66.7 58.8 43.8 71.4 28.6 70.8 50 31.8 22.7 27.8 36.8 63.2 66.7

2.29 (1.15–4.53) 1.94 (0.94–3.98) 1 1.04 (0.43–2.51) 1 2.78 (0.93–8.30) 3.28 (1.09–9.93) 8.56 (1.19–10.64) 2.06 (0.94–4.51) 1.53 (0.64–3.67) 2.5 (1.18–5.31) 1 3.12 (1.38–7.02) 2.2 (0.92–5.29) 1.4 (0.52–3.74) 1 1 1.33 (0.51–3.43) 2.27 (1–5.16) 2.4 (1.06–5.51)

Vegetation size (mm)

INR

Albumin (g/dL)

C-reactive protein (mg/L)

Table III. Logistic regression analysis of potential independent risk factors for embolism Variables*

OR†

95% CI

P

Age Vegetation size C-reactive protein Albumin

0.39 2.09 2.22 0.57

0.21–0.76 1.02–4.27 1.09–4.52 0.29–1.09

.006 .045 .028 .09

*Quartiles have been used as cut off for the analysis, starting from the lower to the higher quartiles. Each variable was adjusted for the confounding effect of all other listed variables and for sex. †OR is the predicted change in odds for a quartile increase in the predictor.

with embolic risk in the univariate analysis. A young age, an increasing vegetation size, and serum CRP levels were the independent variables associated with an increased risk of major embolic events.

Discussion Embolism is a frequent complication of IE and accounts for a substantially higher morbidity and mortality rate.28 The prevalence of embolic events in retrospective studies ranged between 30% and 44%,2,16,20,22,29,30 whereas, in a prospective study over a period of 3 years, 51% of patients underwent ⱖ1 embolic event.3 In our cohort, 46% of patients had a major embolic event during hospitalization. Thus, identifying subgroups of patients with IE at higher risk for embolic events may be crucial for the correct management of the disease.31 At present, early valve replacement is the only effective treatment applied to prevent

P .017

.041

.067

.005

.43

embolic complications in patients with risk of embolism from endocardial vegetations.5 Unfortunately, surgery during the active phase of the disease may predispose to a recurrence of the infection.32 Preliminary results of clinical33 and experimental studies34,35 seem to suggest a beneficial effect of aspirin in reducing the vegetation size, and ongoing studies are evaluating the possible role of aspirin in reducing the rate of embolism in IE. These studies may open new therapeutic perspectives and encourage investigations aimed at evaluating risk factors for embolic events. Our study indicates that a young age, the presence of a large endocardial vegetation and high serum CRP levels are the only factors independently associated with major embolic events during hospitalization. No other factor, such as sex, site of infection, duration of disease, or pathogen involved was independently associated with an increased embolic risk in our IE patients. Our data confirm those reported by others11,12,24,36 that patients with a large vegetation have an increased risk of embolism. Contrasting data have been reported regarding the effect of age on increasing the embolic risk. De Castro et al20 found no statistically significant difference in age between patients with and those without embolic events. Lancellotti et al19 reported that greater age was independently associated with systemic embolization. Furthermore, in a subgroup of patients with native valve disease, it has been reported21 that the clinical presentation, characteristics, and outcome were similar for elderly (aged ⬎64 years) and younger adult patients. Our study is substantially different from those mentioned above;

American Heart Journal Volume 146, Number 2

because it included a larger number of patients with native and prosthetic disease and only major embolic events were assessed. The mechanism(s) underlying the increased tendency of younger subjects to have a major embolic complication during IE is not obvious and requires clarification. We can hypothesize that younger subjects with endocarditis may react more vigorously than older subjects to an inflammatory stimulus, such as bacteremia, and this might predispose to embolism. This interpretation appears to be supported by the fact that lower levels of serum albumin were found in embolic patients, which could indicate a switch in the hepatic protein synthesis towards acute reactants. An association between elevated CRP levels and endocarditis with a complicated course has been described.37 It is known that CRP activates the coagulation cascade by inducing the production of tissue factor by monocytes,38 which can be present in infective endocarditis vegetation. Moreover, because CRP may have an inhibitory effect on platelet aggregation,39 – 41 a higher CRP concentration could modify the platelet function within the vegetation, making it more friable. It should also be mentioned that effective antibiotic treatment, which is known to progressively reduce the embolic rate,5,42 also induces the attenuation of the acute-phase reaction. The difference in the CRP concentration we found between the groups was not mirrored by a similar difference in the fibrinogen or ␣-globulin concentrations. Indeed, discordance regarding the serum levels of different acute-phase proteins is common.43,44 It is important to underline that after an inflammatory stimulus, the CRP levels rise dramatically within a few hours and peak in 2 days.45 The early referral of patients with the features of a more severe condition could have accounted for the difference in the CRP concentration. In our study, the interval between the onset of symptoms and hospital admission was similar in the 2 patient groups and can be ruled out as the reason for the CRP concentration difference. However, we are unable to exclude that the higher inflammatory response in patients who subsequently had a major embolic event was due to microembolic events, which had previously occurred but were not clinically manifest. Like other reports,2,18 our study showed a trend toward a lower frequency of embolic events in patients with endocarditis on prosthetic valves than in those with a native valve infection. Further studies are necessary to clarify this issue. Our patients with prosthetic valves showed a vegetation size smaller than that observed in patients with native valve disease (data not shown). Moreover, all patients with prosthetic infection were receiving warfarin treatment. Whether warfarin could affect the mechanisms of vegetation pro-

Durante Mangoni et al 315

duction, its size, and tendency to originate emboli remains to be evaluated.22,42 In addition, it should be pointed out that subjects with a prosthetic valve are usually older, and indeed in our group the median age was 44 in cases of native valve IE and 52 in those with prosthetic disease (P⬍.05). In conclusion, this study suggests that 3 parameters—a younger age, a larger vegetation size, and higher CRP levels—should be used to identify patients with IE at a higher risk for major embolic events during hospitalization. We thank Mrs Geltrude Fiorillo for valuable technical assistance.

References 1. Bayer AS. Infective endocarditis. Clin Infect Dis 1993;17:313–22. 2. Sandre RM, Shafran SD. Infective endocarditis: review of 135 cases over 9 years. Clin Infect Dis 1996;22:276 – 86. 3. Millaire A, Leroy O, Gaday V, et al. Incidence and prognosis of embolic events and metastatic infections in infective endocarditis. Eur Heart J 1997;18:677– 84. 4. Ting W, Silverman N, Levitsky S. Valve replacement in patients with endocarditis and cerebral septic emboli. Ann Thorac Surg 1991;51:18 –22. 5. Bayer AS, Bolger AF, Taubert KA, et al. Diagnosis and management of infective endocarditis and its complications. Circulation 1998;98:2936 – 48. 6. Jaffe WM, Morgan DE, Pearlman AS, et al. Infective endocarditis, 1983–1988: echocardiographic findings and factors influencing morbidity and mortality. J Am Coll Cardiol 1990;15:1227–33. 7. Roder BL, Wandall DA, Espersen F, et al. Neurologic manifestations in Staphylococcus aureus endocarditis: a review of 260 bacteremic cases in nondrug addicts. Am J Med 1997;102:379 – 86. 8. Reinhartz O, Herrmann M, Redling F, et al. Timing of surgery in patients with acute infective endocarditis. J Cardiovasc Surg 1996; 37:397– 400. 9. Olaison L, Hogevik H, Myken P, et al. Early surgery in infective endocarditis. Q J Med 1996;89:267–78. 10. Le Cam B, Guivaech G, Boles JM, et al. Neurological complications in a group of 86 bacterial endocarditis cases. Eur Heart J 1984;5:97–100. 11. Siddiq S, Missri J, Silverman DI. Endocarditis in an urban hospital in the 1990s. Arch Intern Med 1996;156:2454 – 8. 12. Steckelberg JM, Murphy JG, Ballard D, et al. Emboli in infective endocarditis: the prognostic value of echocardiography. Ann Intern Med 1991;114:635– 40. 13. Sanfilippo AJ, Picard MH, Newell JB, et al. Echocardiographic assessment of patients with infectious endocarditis: prediction of risk for complications. J Am Coll Cardiol 1991;18:1191–9. 14. Buda AJ, Zotz RJ, LeMire MS, et al. Prognostic significance of vegetations detected by two-dimensional echocardiography in infective endocarditis. Am Heart J 1986;112:1291– 6. 15. Stewart JA, Silimperi D, Harris P, et al. Echocardiographic documentation of vegetative lesions in infective endocarditis: clinical implications. Circulation 1980;61:374 – 80. 16. Goldman ME, Fisher EA, Winters S, et al. Early identification of patients with native valve infectious endocarditis at risk for major

316 Durante Mangoni et al

17.

18.

19.

20.

21.

22.

23.

24.

25. 26.

27.

28. 29.

30.

complications by initial clinical presentation and baseline echocardiography. Int J Cardiol 1995;52:257– 64. Rohmann S, Erbel R, Gorge G, et al. Clinical relevance of vegetation localization by transoesophageal echocardiography in infective endocarditis. Eur Heart J 1992;13:446 –52. Mugge A, Daniel WG, Frank G, et al. Echocardiography in infective endocarditis: reassessment of prognostic implications of vegetation size and determined by the transthoracic and the transesophageal approach. J Am Coll Cardiol 1989;14:631– 8. Lancellotti P, Galiuto L, Albert A, et al. Relative value of clinical and transesophageal echocardiographic variables for risk stratification in patients with infective endocarditis. Clin Cardiol 1998; 21:572– 8. De Castro S, Magni G, Beni S, et al. Role of transthoracic and transesophageal echocardiography in predicting embolic events in patients with active infective endocarditis involving native cardiac valves. Am J Cardiol 1997;80:1030 – 4. Gagliardi JP, Nettles RE, McCarty DE, et al. Native valve infective endocarditis in elderly and younger adult patients: comparison of clinical features and outcomes with use of the Duke criteria and the Duke Endocarditis Database. Clin Infect Dis 1998;26:1165– 8. Benn M, Hagelskjaer LH, Tvede M. Infective endocarditis, 1984 through 1993: a clinical and microbiological survey. J Intern Med 1997;242:15–22. Kupferwasser LI, Hafner G, Mohr-Kahaly S, et al. The presence of infection-related antiphospholipid antibodies in infective endocarditis determines a major risk factor for embolic events. J Am Coll Cardiol 1999;33:1365–71. Koie S, Iwase M, Hasegawa K, et al. Echocardiographic prediction of risk for embolism in patients with infective endocarditis. J Cardiol 1997;29(2 Suppl):117–22. Schunemann S, Werner GS, Schulz R, et al. Embolic complications in bacterial endocarditis. Z Kardiol 1997;86:1017–25. Heinle S, Wilderman N, Harrison JK, et al. Value of transthoracic echocardiography in predicting embolic events in active infective endocarditis. Am J Cardiol 1994;74:799 – 801. Durack DT, Lukes AS, Bright DK, et al. New criteria for diagnosis of infective endocarditis: utilization of specific echocardiographic findings: Duke Endocarditis Service. Am J Med 1994;96:200 –9. Scott Harris P, Glenn Cobbs C. Cardiac, cerebral and vascular complications of infective endocarditis. Cardiol Clin 1996;14:437–50. Tornos MP, Olona M, Permanyer-Miralda G, et al. Is the clinical spectrum and prognosis of native valve infective endocarditis in non-addicts changing? Eur Heart J 1995;16:1686 –91. Schulz R, Werner GS, Fuchs JB, et al. Clinical outcome and echocardiographic findings of native and prosthetic valve endocarditis in the 1990s. Eur Heart J 1996;17:281– 8.

American Heart Journal August 2003

31. Erbel R, Liu F, Ge J, et al. Identification of high-risk subgroups in infective endocarditis and the role of echocardiography. Eur Heart J 1995;16:588 – 602. 32. d’Udekem Y, David TE, Feindel CM, et al. Long-term results of surgery for active infective endocarditis. Eur J Cardiothorac Surg 1997;11:46 –52. 33. Taha TH, Durrant SS, Mazeika PK, et al. Aspirin to prevent growth of vegetations and cerebral emboli in infective endocarditis. J Intern Med 1992;231:543– 6. 34. Kupferwasser LI, Yeaman MR, Shapiro SM, et al. Acetylsalicylic acid reduces vegetation bacterial density, hematogenous bacterial dissemination, and frequency of embolic events in experimental Staphylococcus aureus endocarditis through antiplatelet and antibacterial effects. Circulation 1999;99:2791–7. 35. Nicolau DP, Marangos MN, Nightingale CH, et al. Influence of aspirin on development and treatment of experimental Staphylococcus aureus endocarditis. Antimicrob Agents Chemother 1995; 39:1748 –51. 36. Tischler MD, Vaitkus PT. The ability of vegetation size on echocardiography to predict clinical complications: a meta-analysis. J Am Soc Echocardiogr 1997;10:562– 8. 37. Olaison L, Hogevik H, Alestig K. Fever, C-reactive protein, and other acute-phase reactants during treatment of infective endocarditis. Arch Intern Med 1997;157:885–92. 38. Cermak J, Key NS, Bach RR, et al. C-reactive protein induces human peripheral blood monocytes to synthesize tissue factor. Blood 1993;82:513–20. 39. Vigo C. Effect of C-reactive protein on platelet-activating factorinduced platelet aggregation and membrane stabilization. J Biol Chem 1985;260:3418 –22. 40. Fiedel BA, Simpson RM, Gewurz H. Effects of C-reactive protein (CRP) on platelet function. Ann N Y Acad Sci 1982;389:263–73. 41. Fiedel BA. Influence of tuftsin-like synthetic peptides derived from C-reactive protein (CRP) on platelet behaviour. Immunology 1988; 64:487–93. 42. Paschalis C, Pugsley W, John R, et al. Rate of cerebral embolic events in relation to antibiotic and anticoagulant therapy in patients with bacterial endocarditis. Eur Neurol 1990;30:87–9. 43. Gabay C, Kushner I. Acute phase proteins and other systemic responses to inflammation. N Engl J Med 1999;340:448 –54. 44. Mackiewicz A, Sperloff T, Ganapathi MK, et al. Effects of cytokine combinations on acute phase protein production in two human hepatoma cell lines. J Immunol 1991;146:3032–7. 45. Morley JJ, Kushner I. Serum C-reactive protein levels in disease. Ann N Y Acad Sci 1982;389:406 –18.