Skip Navigation

European Journal of Echocardiography 2004 5(6):416-421; doi:10.1016/j.euje.2004.03.005
© 2004 by European Society of Cardiology
This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrowRequest Permissions
Right arrow Disclaimer
Google Scholar
Right arrow Articles by Law, A.
Right arrow Articles by Huynh, T.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Law, A.
Right arrow Articles by Huynh, T.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

Copyright © 2004, The European Society of Cardiology

Negative predictive value of multiplane transesophageal echocardiography in the diagnosis of infective endocarditis

Angeline Lawa, George Honosb and Thao Huynha,*

aDivision of Cardiology, Department of Medicine, McGill University Health Center, McGill University, Canada
bDivision of Cardiology, Department of Medicine, Sir Mortimer B. Davis, Jewish General Hospital, McGill University, Canada

Received 4 December 2003; received in revised form 8 March 2004; accepted after revision 11 March 2004.

* Corresponding author. Montreal General Hospital, McGill University Health Center, Division of Cardiology, 1650 Cedar Avenue, Room #E5-200, Montreal, Quebec, Canada PQ H3G 1A4. Tel.: +1-514-934-1934x43241; fax: +1-514-934-8318. thao.huynh{at}muhc.mcgill.ca


   Abstract
Top
 Abstract
Background
 Methods
 Results
 Discussion
 Limitations
 Conclusions
 Notes
 References
 
Background: The clinical implications of a negative multiplane transesophageal echocardiography (TEE) have not yet been reported. We aim to determine the negative predictive value (NPV) of a negative multiplane TEE in patients with suspected infective endocarditis (IE).

Methods and results: We identified 83 consecutive patients with suspected IE and negative multiplane TEE from our echocardiographic database. Of 74 patients with a minimum of 1-month follow-up, only 1 patient developed "definite IE". Eight patients had "possible IE". The calculated NPV of multiplane TEE in IE was 98.6% if we only considered the case of "definite IE". If we assumed that all patients with "possible IE" had the disease, then the NPV of multiplane TEE was 87.8%.

Conclusions: Multiplane TEE is a highly accurate diagnostic tool with excellent NPV in IE. However, in a highly suspicious clinical setting for IE, a repeat TEE is still recommended to assess evolving echocardiographic features.

Keywords: Echocardiography; Transesophageal echocardiography; Infective endocarditis


   Background
Top
 Abstract
 Background
Methods
 Results
 Discussion
 Limitations
 Conclusions
 Notes
 References
 
Echocardiography plays an important role in establishing the diagnosis of infective endocarditis.1–16 The negative predictive value (NPV) of monoplane TEE has been previously reported as 86–92%.17 Using both monoplane and biplane TEEs, Lowry et al. documented an NPV of 100% in native valves and 90% in prosthetic valves.18 The introduction of multiplane TEE16 has further enhanced the detailed visualization of cardiac structures. Although there had been no direct comparison, multiplane TEE should be superior to single plane or biplane TEE for the detection of small vegetations or an abscess. However, the clinical impact of a negative multiplane TEE has not been reported. Can a negative multiplane TEE in patients with suspected IE exclude this diagnosis with certainty? The primary objective of this study was therefore to determine the NPV of a multiplane TEE in patients with suspected IE.


   Methods
Top
 Abstract
 Background
 Methods
Results
 Discussion
 Limitations
 Conclusions
 Notes
 References
 
We searched the echocardiography databases of 2 university teaching hospitals for patients with a negative TEE for suspected infective endocarditis, from January 1997 to December 2000. All studies were performed with a multiplane 5 MHz phased-array TEE probe using one of the following echocardiography systems: Phillips/Agilent 2500 or 5500, Siemens/Acuson Sequoia C256 or General Electric/Vivid 5. All TEEs were performed by level III echocardiographers.

We excluded all patients without a minimum follow-up of 1 month. We compiled data on clinical presentation, hospital course, blood culture results, and pathological confirmation of IE, surgical intervention, discharge diagnosis, and clinical outcomes. We also recorded predisposing factors for IE including prior history of IE, intravenous drug abuse, recent dental work, rheumatic heart disease, and presence of indwelling catheter.

Echocardiographic criteria
TEE was considered negative for IE in the absence of the following features suggestive of IE, as defined by the Duke criteria:19

1. Vegetation defined as an oscillating intracardiac mass on valve or its supporting structures, in the path of regurgitant jets, or on implanted material in the absence of an alternative anatomical explanation.
2. An abscess was defined as an echo-free space in the paravalvular area or adjacent structures.
3. New dehiscence of a prosthetic valve.

Follow-up
We reviewed all charts for physician's follow-up examinations, as well as surgical, pathology and autopsy reports. Time intervals were determined to the nearest month.

Classification of infective endocarditis
We used both the proposed modified Duke criteria20,21 and final physician assessment, to determine whether a patient had "definite IE" or "possible IE". Patients were diagnosed as "possible IE" only if confirmed by both the final discharge diagnosis and the modified Duke criteria. Case classification was done by 2 independent reviewers.

The original and modified Duke criteria have the same requirements for "definite IE": 2 major criteria or 1 major and 3 minor criteria or 5 minor criteria. However, requirements differ for the diagnosis of "possible IE", with the original Duke criteria requiring a minimum of only 1 minor criterion for IE be met to exclude a diagnosis of "rejected IE". Thus, a patient could be diagnosed with the original Duke criteria as "possible IE" with only 1 minor criterion met. The modified Duke criteria require a patient to have at least 1 major and 1 minor criteria or 3 minor criteria to qualify as "possible IE".

Statistical analysis
Negative predictive value was defined as the proportion of true negatives among the negatives (true negatives/(true negatives + false negatives). Unless otherwise specified, all data were expressed as mean value ± SD. Statistical analysis was performed using SPSS 11.0 for Windows, Chicago.


   Results
Top
 Abstract
 Background
 Methods
 Results
Discussion
 Limitations
 Conclusions
 Notes
 References
 
We identified 83 patients with negative TEE for IE. We excluded 9 patients for lack of a minimum 1-month follow-up. Clinical characteristics for the 74 remaining patients are shown in Table 1. There were 40 men and the mean age was 64.5 ± 15 years (range 19–88). Almost half of the patients (44.5%) were elderly (defined as older than 65 years).


View this table:
[in this window]
[in a new window]

  		Table 1 Clinical variables

 
Risk factors for development of IE are shown in Table 1. Twenty-seven patients (36.4%) had predisposing conditions for IE, which included prosthetic valves (21.6%), prior IE (8.1%), illicit intravenous drug use (4.0%), congenital intracardiac malformations (5.4%), degenerative (5.4%) and rheumatic valvular diseases (2.7%). Sixteen patients had prosthetic valves: 15 metallic (7 aortic, 8 mitral), 6 bioprosthesis (3 aortic, 2 mitral and 1 pulmonary homograft valve). Ten patients (13.5%) had intravascular catheters, which included 8 hemodialysis catheters, 1 Port-a-Cath, and 1 Broviac.

The majority of patients (71.6%) developed fever in the days prior to their TEE whereas only a small number were noted to have a new or previously not noted heart murmur (11.0%). One patient(1.3%) had an elevated rheumatoid factor. Three patients (4.1%) had vascular phenomena with possible Janeway's lesions. Thirty-nine patients (52.7%) had documented bacteremia during their admission. Two patients had polymicrobial bacteremia. The responsible microorganisms are shown in Fig. 1.


Figure 1
View larger version (22K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Figure 1 Organisms responsible for bacteremia.

 
Table 2 shows the final discharge diagnoses. The 3 most frequent discharge diagnoses were sepsis, pneumonia, and fever not yet diagnosed. There were 14 deaths (18.9%) during the hospitalization; none could be attributed as a direct consequence of IE (Table 3). However, none of these patients had an autopsy. Seven patients were treated and discharged with the final diagnosis of "possible IE" based on clinical symptoms, bacteremia with typical pathogens, good clinical response defined as resolution of fever and symptoms with appropriate antibiotic therapy, and no other identified source of bacteremia/infection (Table 4). All fulfilled the modified Duke criteria of "possible IE". None developed complications of IE or had pathological confirmation of IE.


View this table:
[in this window]
[in a new window]

  		Table 2 Final diagnoses of patients without infective endocarditis

 


View this table:
[in this window]
[in a new window]

  		Table 3 Causes of deaths

 


View this table:
[in this window]
[in a new window]

  		Table 4 Clinical characteristics of patients classified and treated as "possible infective endocarditis"

 
The mean follow-up was 7.9 ± 6 months. At least 3-month follow-up data were available in 62 patients (83.7%). Forty patients (54.5%) had a minimum of 6-month follow-up. Only 1 patient required repeat TEE at 3 months following discharge. During the index hospitalization, no patients required repeated TEE.

There were 7 late deaths. All except one occurred more than 12 months following the initial discharge. The early death was due to an acute myocardial infarction at 6 weeks. None of the late deaths could be attributed to IE.

Only 1 patient developed "definite IE". She was a 42-year-old woman with a bicuspid aortic valve admitted with fever, 1 set of blood cultures positive for group B streptococcus, and a negative TEE study. She was treated with ampicillin and gentamicin and discharged home on oral penicillin with the presumptive diagnosis of reactive polyarthritis and urinary tract infection. At 3 months, she presented with 3-week history of progressive dyspnea. Repeated TEE revealed a vegetation on the aortic valve and new severe aortic regurgitation. All blood cultures remained negative during the second hospitalization. The patient was started empirically on parenteral penicillin G. She underwent an aortic valve replacement and concomitant closure of an atrial septal defect with a pericardial patch. At surgery, there was a perforation of an aortic valve leaflet. The surgeon did not detect any vegetation or abscess. Unfortunately, the excised aortic valve was not sent for cultures or pathology. At 7 months following this surgery, the patient was doing well.

The negative predictive value of a multiplane TEE was 73/74 = 98.6%, when we only accounted for the case of "definite IE". If we assumed that all of the 7 cases of "possible IE" truly had the disease, then the NPV of the multiplane TEE in IE would be (74 – (8 possible IE + 1 definite IE))/74 = 87.8%.


   Discussion
Top
 Abstract
 Background
 Methods
 Results
 Discussion
Limitations
 Conclusions
 Notes
 References
 
Roe et al.13 studied the incremental value of TEE in patients with suspected IE, however, they focused mainly on the positive predictive value of TEE in incremental value to the original Duke criteria rather than the negative predictive value. Our study was the first to incorporate both the modified Duke criteria and multiplane TEE in the diagnosis of IE. We preferred to rely on firm diagnostic criteria with objective, impartial case evaluation rather than on subjective physician assessment, as in previous studies.16,17,22 We selected the modified Duke criteria rather than the original Duke criteria because the proposed modifications are stricter for the diagnosis of "possible IE".19,20

Our patients appeared to be at intermediate to high risk for developing IE, with 37% having predisposing conditions. The majority of these patients had fever and bacteremia. Hospital mortality was high (18.9%), reflecting the severity of illness. Ten deaths were due to sepsis. Infective endocarditis could not be confirmed in any of these patients. TEE was particularly useful in the 39 patients with bacteremia. A negative multiplane TEE was helpful to exclude IE in 35 patients. Four of the remaining patients received treatment as "possible IE". The NPV of the multiplane TEE in this subgroup of patients was 35/39 = 89.7%.

Multiplane TEE was highly accurate in all patients with prosthetic valves. None of these patients developed definite IE, although 1 patient had bacteremia and was treated as "possible IE". The calculated NPV in patients with prosthetic valves was 15/16 = 93.8%.

Single plane TEE has been reported to have NPV of 86%–92%.17 Using both single and biplane TEE, Lowry et al.18 documented NPV of 90% in prosthetic valves and 100% in native valves. Although introduced since 1992,16 the clinical implications of a multiplane TEE in IE have not been previously studied. Our calculated NPV of multiplane TEE was 98.6% if only the case of "definite IE" was taken into account. However, if we assumed that all patients with "possible IE" had the disease, then the NPV of multiplane TEE would be 87.8%. This still compared favorably with the NPV of single plane TEE reported by Sochowsky and Chan17 in similar group of patients treated as IE without definite confirmation.

Our single patient with "definite IE" had vegetation and new aortic insufficiency on the repeat TEE. Although no vegetation was found at surgery, it was possible that it may have embolized. It was unfortunate that we did not have pathological confirmation in this case. However, the valvular perforation documented at surgery, was likely a direct result of IE. Of note, this patient had IE involving a native valve. The failure to detect vegetation in her case was likely secondary to the fact that TEE was performed too early in the course of the disease (1 day after the onset of the symptoms). The vegetation may be exceedingly small to be able to be detected even by multiplane TEE, or may not even be there. This finding underscored the importance of repeating TEE in a highly suspicious clinical context. Although Lowry et al.18 reported an NPV of 100% of TEE in native valves, their mean follow-up was only 25.4 ± 23 days for 40 patients. Infective endocarditis was recognized in our patient, only at 3 months after the initial bacteremia. This fact emphasized the importance of adequate follow-up interval for patients with suspected IE.


   Limitations
Top
 Abstract
 Background
 Methods
 Results
 Discussion
 Limitations
Conclusions
 Notes
 References
 
The minimum follow-up was only 1 month. However, 83.7% of our patients had at least 3-month follow-up with a mean time interval of 7.9 ± 6 months. In our practice, patients originally evaluated at our institution are almost always transferred back to our institution if IE subsequently develops. It would be unlikely that any of the 12 patients, with missing 3-month follow-up, would have developed IE without our knowledge.

It was possible that some of the deaths were due to missed IE, since we did not have any post-mortem examinations for the in-hospital deaths. However, all of the in-hospital deaths had a clearly identified alternate etiology. All except one of the out-of hospital deaths occurred more than 12 months later. It would be extremely unusual that missed IE remained unrecognized for such a long time interval.


   Conclusions
Top
 Abstract
 Background
 Methods
 Results
 Discussion
 Limitations
 Conclusions
Notes
 References
 
Multiplane TEE is an important clinical tool in the diagnosis of IE. Its NPV ranged from 87.8% to 98.6%. The NPV of TEE remained excellent in high-risk patients such as those with prosthetic valves and those with bacteremia (93.% and 89.7%, respectively).

However, although it is highly accurate, multiplane TEE can still miss rare cases of IE, especially if performed too early in the course of the disease as vegetations may be too small to be detected at that stage. In highly suspicious cases, such as patients with bacteremia from typical micro-organisms, a follow-up TEE is indicated to assess for any new vegetation or other complications of IE.3,17 Good clinical judgment and adequate long-term follow-up are therefore essential, as IE is a challenging diagnosis which may remain unrecognized for a prolonged period of time.


   Notes
Top
 Abstract
 Background
 Methods
 Results
 Discussion
 Limitations
 Conclusions
 Notes
References
 
{star} Part of this manuscript was presented at Canadian Cardiovascular Society Congress 2001, Halifax, Canada.


   References
Top
 Abstract
 Background
 Methods
 Results
 Discussion
 Limitations
 Conclusions
 Notes
 References
 

  1. Bayer A.S., Bolger A.F., Taubert K.A., Wilson W., Steckelberg J., Karchmer A., et al. Diagnosis and management of infective endocarditis and its complication. Circulation (1998) 98:2936–2948.[Free Full Text]
  2. Mylonakis E., Calderwood S.B. Infective endocarditis in adults. N Engl J Med (2001) 345(18):1318–1330.[Free Full Text]
  3. Yvorchuk K.J., Chan K.L. Application of transthoracic and tranesophageal echocardiography in the diagnosis and management of infective endocarditis. J Am Soc Echocardiogr (1994) 14:294–308.
  4. Erbel R., Rohmann S., Drexler M., Mohr-Kahaly S., Gerharz C.D., Iversen S., et al. Improved diagnostic value of echocardiography in patients with infective endocarditis by transoesophageal approach: a prospective study. Eur Heart J (1988) 9:43–53.[Abstract/Free Full Text]
  5. Daniel W.G., Schroeder E., Nonnast-Daniel B., Litchtlen P.R., et al. Conventional and transesophageal echocardiography in the diagnosis of infective endocarditis. Eur Heart J (1987) 8(Suppl. J):287–292.[Abstract/Free Full Text]
  6. Mugge A. Echocardiographic detection of cardiac valve vegetations and prognostic implications. Infect Dis Clin North Am (1993) 7(4):877–898.[Web of Science][Medline]
  7. Birmingham G.D., Rahko P.S., Ballantyne F. Improved detection of infective endocarditis with transesophageal echocardiography. Am Heart J (1992) 123:774–781.[CrossRef][Web of Science][Medline]
  8. Shively B.K., Gurule F.T., Rohadan C.A., Leggett J.M., Schiller N.B. Diagnostic value of transesophageal compared with transthoracic echocardiography in infective endocarditis. J Am Coll Cardiol (1991) 18:391–397.[Abstract]
  9. Thalme A., Nygren A.T., Julander I., Freyschuss U. Endocarditis: clinical outcome and benefit of trans-oesophageal echocardiography. Scand J Infect Dis (2000) 32(3):303–307.[CrossRef][Web of Science][Medline]
  10. Daniel W.G., Mugge A., Martin R.P., Lindert O., Haussman D., Nonnast-Daniel B., et al. Improvement in the diagnosis of abscesses associated with endocarditis by transesophageal echocardiography. N Engl J Med (1991) 324:795–800.[Abstract]
  11. Karalis D.G., Bansal R.C., Hauck A.J., Ross J.J. Jr., Applegate P.M., Jutzy K.R., et al. Transesophageal echocardiographic recognition of subaortic complications in aortic valve endocarditis: clinical and surgical implications. Circulation (1992) 86:353–362.[Abstract/Free Full Text]
  12. De Castro S., Carton D., D'Amati G., Beni S., Yao J., Fiozell M., et al. Diagnostic accuracy of transthoracic and multiplane transesophageal echocardiography for valvular perforation in acute infective endocarditis: correlation with anatomic findings. Clin Infect Dis (2000) 30(5):825–826.[CrossRef][Web of Science][Medline]
  13. Roe M.T., Abramson M.A., Heinle S.K., Kisslo J., Corey G.R., Sexton D.J. Clinical information determines the impact of transesophageal echocardiography on the diagnosis of infective endocarditis by the Duke criteria. Am Heart J (2000) 139(6):945–951.[CrossRef][Web of Science][Medline]
  14. Lengyel M. The impact of transesophageal echocardiography on the management of prosthetic valve endocarditis: experience of 31 cases and review of the literature. J Heart Valve Dis (1997) 6(2):204–211.[Web of Science][Medline]
  15. Schulz R., Werner G.S., Fuchs J.B., Andreas S., Prange H., Ruschewski W., et al. Clinical outcome and echocardiographic findings of native and prosthetic valve endocarditis in the 1990s. Eur Heart J (1996) 17:281–288.[Abstract/Free Full Text]
  16. Sochowski R.A., Chan K.L. Implication of negative results on a monoplane transesophageal echocardiographic study in patients with suspected infective endocarditis. J Am Coll Cardiol (1993) 21:216–221.[Abstract]
  17. Lowry R., Zoghbi W.A., Baker W.B., Wray R.A., Quinones M.A. Clinical impact of transesophageal echocardiography in the diagnosis and management of infective endocarditis. Am J Cardiol (1994) 73:1089–1091.[CrossRef][Web of Science][Medline]
  18. Pandian N.G., Hsu T.L., Schwartz S.L. Multiplane transesophageal echocardiography: imaging planes, echocardiographic anatomy, and clinical experience with a prototype phased array multiplane probe. Echocardiography (1992) 9:649–666.[Web of Science][Medline]
  19. Durack DT, Lukes AS, Bright DK, and the Duke Endocarditis Service. New criteria for diagnosis of infective endocarditis: utilization of specific echocardiographic findings. Am J Med 1994;96:200–9.
  20. Li J.S., Sexton D.J., Mick N., Nettles R., Fouler V.G., Ryan T., et al. Proposed modifications to the Duke criteria for the diagnosis of infective endocarditis. Clin Infect Dis (2000) 30:633–638.[CrossRef][Web of Science][Medline]
  21. Bayer A. Revised diagnostic criteria for infective endocarditis. Cardiol Clinics (1996) 14(3):345–350.[CrossRef][Medline]
  22. Lindner J.R., Case A., Dent J.M., Abott R.D., Scheld W.M., Kaul S. Diagnostic value of echocardiography in suspected endocarditis, an evaluation based on the pretest probability of disease. Circulation (1996) 93:730–736.[Abstract/Free Full Text]

Add to CiteULike CiteULike   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us    What's this?



This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrowRequest Permissions
Right arrow Disclaimer
Google Scholar
Right arrow Articles by Law, A.
Right arrow Articles by Huynh, T.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Law, A.
Right arrow Articles by Huynh, T.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?