1

Infective endocarditis

Learning objectives

When you have completed this you should understand:

  1. The diagnosis of infective endocarditis
  2. The role of investigationsincluding echocardiography in infectiveendocarditis
  3. Organismsassociated with infective endocarditis
  4. Rational antibiotictreatment of patients with infective endocarditis
  5. The significanceof large vegetations

A 70 year old man presented with microscopic haematuria andproteinuria and a fever five months after having a transurethralresection of the prostate (TURP). Initially urological reviewwas arranged as the family doctor thought that a urinary infectionwas the most likely diagnosis.

The patient was concerned that he was not getting better andhe self-referred to a physician. He had continuing fever, weightloss, and malaise. The physician detected a mitral pan-systolicmurmur that had not been heard before. On the basis of thisfinding infective endocarditis was suspected and investigationsbegun. His subsequent course and its management are discussedin an interactive case presentation.

Case History

  • 70 year old man presented with a three month history of malaise,lethargy, and 8 kg weight loss. Five months previously he hada transurethral resection of the prostate (TURP) for benignprostatic hypertrophy. The early postoperative period was complicatedby retention of urine requiring reinsertion of a temporary catheter.
  • On examination there was a pansystolic murmur consistent withmitral regurgitation. The apex beat was not displaced and noother cardiac abnormality was documented. The patient had atemperature of 37°C

Table 1 Results of routine investigations

Haematology / Biochemistry
Hb(g/dl) / 11.3 / Na (mmol/l) / 130
WCC (x 109/l) / 11.5 / K (mmol/l) / 4.9
Platelets(x 109l) / 490 / Urea (mmol/l) / 5.9
MCV (fl) / 83 / Creatinine (mmol/l) / 109

Hb,haemoglobin; MCV, mean cell volume; WCC, white cell count

Liver function tests were also normal.

Table 2 Inflammatory markers and urinalysis

CRP (mg/l) / 110
ESR(mm/1st hour) / 85
Urinalysis
Blood / ++MSU: no growth
Protein / ++

CRP, C reactive protein;ESR, erythrocyte sedimentation rate; MSU, midstream urine

Question 1:

  1. The patient probably has endocarditis and must be started on antibiotics while the results of blood cultures are awaited

THE ANSWER IS F:
There is insufficient information for the diagnosis of endocarditis according to the modified Duke criteria. As the patient is clinically stable, it would be reasonable to wait for the results of blood cultures and organise an echocardiogram before starting antibiotics.

  1. A white cell count of 11.5 x 109/l is unusually low for a patient with active infective endocarditis

THE ANSWER IS F:
The white cell count is usually only mildly raised and can be normal unless the infecting organism is a virulent one such as Staphylococcus aureus.

  1. The most important investigation in the diagnosis of endocarditis is the blood culture

THE ANSWER IS T:
The blood culture is fundamental for the diagnosis of endocarditis and for planning appropriate antibiotic treatment.

  1. Haematuria is an uncommon finding in endocarditis

THE ANSWER IS F:
Haematuria and proteinuria are extremely common findings in endocarditis and are usually caused by the immune glomerulonephritis mediated by the circulating immunoglobulins stimulated by the infection. This may be present even when the urea and creatinine are normal. Urinalysis should be part of the routine assessment of patients with endocarditis and is a simple non-invasive test for monitoring progress. Persistent or recurrent haematuria may indicate failure of medical treatment.

  1. An elevated C reactive protein (CRP) and erythrocyte sedimentation rate (ESR) are required for the diagnosis of endocarditis

THE ANSWER IS F:
The CRP and ESR are not part of the Duke criteria for the diagnosis of endocarditis. However, they are nearly always raised in endocarditis and they are useful for monitoring response to treatment.

Начало формы

The diagnosis of infective endocarditis: The Duke criteria

The diagnosis of infective endocarditis is often not straightforwardand various attempts have been made to standardise it. The bestknown of these attempts is the development of criteria by DukeUniversity in the USA.

In 1994, a group at Duke University proposed standardisedcriteria (Table 3) for assessing patients with suspected infectiveendocarditis.1 These criteria have proved useful and are widelyused.

Table 3 The Duke criteria
Major criteria:
Positive bloodcultures of typical bacteria
  • Viridans streptococci
  • Streptococcusbovis
  • Organisms from HACEK group
  • Staphylococcus aureus
  • Enterococci
Evidenceof endocardial infection
  • Vegetation, abscess or prostheticvalve dehiscence on echocardiography
  • New valvar regurgitationon echocardiography
Minor criteria:
Predisposing heartcondition or intravenous drug abuse
  • Fever >38°C
  • Embolic/vascularphenomena
  • Immunological phenomena (e.g. glomerular nephritis,rheumatoid factor, Osler's nodes, Roth spots)
  • Serology consistentwith infective endocarditis
  • Blood cultures compatible but nottypical for endocarditis
  • Other echo findings consistent withinfective endocarditis and not covered by major criteria
HACEK,Haemophilus species (not influenzae), Actinobacillus actinomycetemcomitans,Cardiobacterium hominis, Eikenella corrodens, Kingella kingae
Reproduced from Durack et al.1 with permission.

Diagnostic clinical parameters and echocardiographic findings of Duke criteria

The Duke criteria combine important diagnostic clinical parameterswith echocardiographic findings (Table 4).

Table 4 Diagnostic clinical parameters and echocardiographicfindings of Duke criteria
Clinical parameters:
  • Persistentbacteraemia
  • New regurgitant murmurs
  • Vascular complications
  • Intravenousdrug use
Echocardiographic findings:
  • Mobile echo dense massesattached to valve leaflets or endocardium
  • Periannular abscesses
  • Newdehiscence of a prosthetic valve
  • New valvar regurgitation

The Duke criteria also includes minor criteria (Table 3, previouspage) that are associated with infective endocarditis but donot have the same diagnostic importance as the major criteria.

CRP and ESR are not part of the Duke criteria; however, theyare usually raised in infective endocarditis and have been proposedas additional minor criteria to improve sensitivity.2

Diagnosis using the Duke criteria

The Duke criteria are used in combination to decide whetherthe diagnosis is infective endocarditis or not (Table 5).

Table 5 Diagnosis using the Duke criteria
Definite:
  • Clinicaldiagnosis is made by 2 major, 1 major and 3 minor, or 5 minorcriteria
  • Pathological diagnosis is made by histological evidenceof active infective endocarditis (from surgical or post-mortemspecimens)
Possible:
  • Findings consistent with infectiveendocarditis but do not satisfy the requirements of definiteinfective endocarditis
Rejected:
  • Firm alternative diagnosis
  • Resolutionof manifestations within 4 days of antibiotic treatment
  • Noevidence of infective endocarditis at surgery or post-mortemafter therapy for >4 days

Sensitivity and specificity

The usefulness of the Duke criteria has been validated inseveral series.3,4,5,6 The calculated negative predictive valueof the Duke criteria was 98% when 52 "rejected" cases of infectiveendocarditis were followed up for longer than three months fora potential "missed diagnosis".4 Another study looked at 100patients with pyrexia of unknown origin who had blood culturesand echocardiograms. They found the Duke criteria to be 99%specific at rejecting endocarditis in patients with unexplainedfever.5

The Duke criteria also has good (72–90%) agreement withclinical assessment by infectious disease experts,6 indicatingit also has reasonable sensitivity for the diagnosis of infectiveendocarditis.

Negative blood cultures

Approximately 5% of cases will satisfy the diagnostic criteriafor infective endocarditis but will have negative blood cultures.This is often the result of prior antibiotic treatment, whichemphasises the importance of taking blood cultures in high riskpatients before treatment with antibiotics. Another cause ofnegative cultures is infection with fastidious organisms.

Fastidious organisms include the HACEK group (Haemophilusspecies (not influenzae), Actinobacillus actinomycetemcomitans,Cardiobacterium hominis, Eikenella corrodens, Kingella kingae).If the clinical suspicion is high for infective endocarditisthen all blood cultures should have prolonged culture to increasethe detection of these fastidious organisms. Serology shouldbe sent for antibodies against Coxiella burnetti, Chlamydia,and Bartonella, which are not readily cultured in most microbiologicallaboratories.

Polymerase chain reaction has also been used to identify certainrare fastidious organisms such as Tropheyma whipplei and Bartonellaspecies.3

Because of the importance of accurate bacteriological advicefor both diagnosis and management, microbiologists should alwaysbe consulted, especially when resistant or atypical organismshave been identified.

The patient was admitted to hospital with a provisional diagnosisof endocarditis. Six out of six blood culture bottles grew Gram-positivestreptococci (fig 1). The organism was confirmed to be Enterococcusfaecalis. It was fully sensitive to amoxycillin but had highlevel resistance to gentamicin and streptomycin.


Figure 1 Gram stain from bloodculture. Gram-positive cocci in pairs and chains

Question 2:

  1. Enterococcus is very rarely the pathogen in patients with endocarditis

THE ANSWER IS F:
Enterococci are the causative pathogens in approximately 10% of cases

  1. Enterococci are part of the normal gut flora

THE ANSWER IS T

  1. If the blood cultures had been positive for viridans streptococci this can almost always be considered diagnostic of endocarditis

THE ANSWER IS T:
Patients who have blood cultures positive for viridans streptococci nearly always have endocarditis. It is extremely uncommon for viridans streptococci to cause foci of infection outside of the heart

  1. If a patient is clinically septic on presentation it is unlikely that they have endocarditis

THE ANSWER IS F:
Patients with Staphylococcus aureus endocarditis are often clinically septic on presentation and the underlying diagnosis of endocarditis may not be apparent on presentation

  1. If endocarditis is suspected clinically but blood cultures remain negative after 48 hours then an alternative diagnosis should be sought

THE ANSWER IS F:
Although most patients with endocarditis have blood cultures that are positive at 48 hours, a significant proportion do not. This may be due to previous administration of antibiotics or alternatively there may be a fastidious organism such as from the HACEK group

Clinical data continued

The patient's ECG on admission is shown in (fig 2).


Figure 2 12 lead ECG on admission:sinus rhythm with left axis deviation

The chest xray was normal. However, a transthoracic echocardiogramsuggested that there may be a vegetation on the mitral valve,but this was not definite because the quality of the imageswere poor for technical reasons.

Question 3:

  1. All patients with suspected endocarditis and a negative transthoracic echocardiogram (TTE) study should undergo a transoesophageal echocardiogram (TOE, or TEE in the USA)

THE ANSWER IS T:
TOE has greater sensitivity then TTE for picking up abscesses and problems with prosthetic valves—for example, valve ring dehiscence. It also provides valuable information about the severity of regurgitant lesions and more detailed information about the nature of vegetations (size, mobility, etc).

  1. A normal TOE scan cannot rule out the diagnosis in a patient with clinically suspected endocarditis

THE ANSWER IS T:
A normal TOE has a negative predictive value of over 90% in excluding endocarditis in clinically suspected cases with a high index of suspicion. However, a single TOE cannot exclude endocarditis. Repeating the TOE after a few days provides further reassurance and is probably the most effective way of excluding endocarditis (negative predictive power of over 95%).

  1. Once infective endocarditis has been diagnosed a TTE need be repeated only when clinically indicated

THE ANSWER IS F:
It is good practice to repeat the echocardiogram in the early stages of antibiotic treatment to monitor progress and look for complications—for example, changes in the severity of regurgitation, structural damage, vegetation size, etc.

  1. Ideally all patients with fever should have an echocardiogram to rule out infective endocarditis

THE ANSWER IS F:
Fever is a very common problem yet endocarditis is a very rare disease. To put this in perspective the condition is sufficiently rare that the average general practitioner (family physician) will not see a single new case during a working lifetime. Thus the pre-test probability of endocarditis needs to be taken into account before proceeding to echocardiography as the diagnostic yield is likely to be very low and the number of false positives high, and the test is expensive and time consuming to perform.

  1. The left axis deviation on the 12 lead ECG means that there is a significant risk of an aortic root abscess involving the conduction system

THE ANSWER IS F:
The left axis deviation has little or no significance. However, the resting 12 lead ECG has definite value in the assessment of patients with infective endocarditis and should be repeated regularly during treatment. Progressively increasing prolongation of the PR interval is an ominous sign as it strongly suggests an abscess in the interventricular septum with involvement of the atrioventricular node—a situation that needs surgical treatment without delay.

Learning Point I: Role of transoesophageal echocardiography in infective endocarditis

  • Transoesophageal echocardiography (TOE) is a safe investigationand extremely useful in the diagnosis and management of patientswith infective endocarditis.
  • TOE has sensitivity and specificityof approximately 90% for the detection of vegetations, comparedto 70% with transthoracic echocardiography (TTE).7,8
  • TOE iseven more useful for the detection of abscesses and prostheticvalve complications. The TOE probe in the oesophagus is in physicalproximity to the aortic root and basal septum where most suchcomplications occur.9 TOE is also much better at assessing prostheticvalves, because there is less acoustic shadowing via the TOEwindow assessment. It is therefore very useful for detectingvegetations, paraprosthetic leaks, and valvar regurgitation.10
  • TOEhas an extremely good negative predictive value of over 90%in excluding infective endocarditis in cases with a high indexof suspicion. However, a single TOE cannot exclude infectiveendocarditis. Repeating the TOE after a few days provides furtherreassurance and is probably the most effective way of excludingendocarditis with a negative predictive power of over 95% 11
  • Fortricuspid valve endocarditis the TTE may actually provide equivalentimage quality (if not better), as the right heart lies closerto the chest wall.
  • Endocarditis antibiotic prophylaxis is currentlynot recommended for TOE.12

Clinical data continued

The patient was transferred to the care of the cardiologyteam and was started on intravenous amoxycillin.

Question 4:

  1. Providing the patient responds appropriately after antibiotics are started he would be suitable for a two week course of antibiotics

THE ANSWER IS F:
Although in selected cases short courses of antibiotics are appropriate, in this case this would not be indicated. Short (two week) courses of antibiotics are usually reserved for patients infected with viridans streptococci providing they meet criteria that put them into a low risk group.

  1. Vancomycin should be added to the amoxycillin as the organism is resistant to both gentamicin and streptomycin

THE ANSWER IS F:
Both amoxycillin and vancomycin work by bacterial cell wall synthesis inhibition. There is no synergy and hence no advantage to add vancomycin to amoxycillin.

  1. A cephalosporin such as cefotaxime is unlikely to be effective as an alternative to penicillin

THE ANSWER IS T:
All enterococci are resistant to cephalosporins.

  1. The patient should be anticoagulated to decrease the risk of embolic complications

THE ANSWER IS F:
There is no evidence that anticoagulation decreases the risk of emboli in endocarditis and most authorities believe that it is contraindicated in patients with native valve endocarditis, who are in sinus rhythm, because of the risk of intracerebral haemorrhage. If the patient requires anticoagulation for some other reason—for example, atrial fibrillation—then anticoagulation should be continued.

  1. The risk of emboli drops significantly after the first week of treatment

THE ANSWER IS T:
The risk of emboli is highest in the first week of antibiotic treatment and drops dramatically thereafter.

Bacteriology and antibiotic treatment of infective endocarditis

General points of importance
  • Historically viridans streptococciwere the most common infecting agents but the microbiologicalprofile of infective endocarditis is changing.
  • Staphylococcalinfections are becoming increasingly common due to the increaseof intravenous drug use, prosthetic valves, and infected intravascularlines.3 Native valve infective endocarditis is now as likelyto be caused by staphylococcus as it is by viridans streptococci.13
  • Itis fundamentally important that patients should always haveblood cultures taken before commencing antibiotic

Starting Treatment

A delay in starting treatment in patients with infective endocarditiscan result in irreversible cardiac damage. However, many patientswill have been unwell for a prolonged period of time and sodelaying treatment for 48 hours until the results of blood culturesare known is not unreasonable in the clinically stable patientwhen the diagnosis is uncertain. However when the diagnosisis obvious or when the patient is septic and generally unwell,as is often the case with staphylococcal infection, then treatmentshould be started as soon as the blood cultures have been taken.

  • The recommendations given here are from the British Societyfor Antimicrobial Chemotherapy.14 They are only guidelines andeach case should be managed on an individual basis with closeliaison with both microbiologists and surgeons.
  • For most streptococci,staphylococci, and enterococci the combinationof penicillinand an aminoglycoside (usually gentamicin) isthe most appropriatetreatment
  • Both penicillins and aminoglycosides are bactericidaland consequentlythey are extremely effective antibiotics fortreating sensitiveorganisms.

Is vancomycin a more potent antibiotic than penicillin forthe treatment of infective endocarditis?

It is often mistakenly believed that vancomycin is a moreeffective antibiotic than penicillin. Switching a patient tovancomycin instead of penicillin when they are not respondingis unlikely to be helpful. When patients fail to respond toappropriate antibiotic treatment they often require surgeryrather than a different antibiotic.

Blind first line treatment

In most patients with infective endocarditis treatment hasto begin before the organism and its sensitivity are known.It is important to choose a combination of treatment that islikely to be effective and then modify the treatment when theresults of blood cultures are available.

  • Intravenous benzyl penicillin and gentamicin are normallyrecommendedas first line antibiotics as they will be activeagainst penicillinsensitive viridans streptococci and Streptococcusbovis, themost common organisms in native valve endocarditis.
  • If there is a strong possibility of staphylococcal infection—forexample, in intravenous drug users, patients on haemodialysis,or who have undergone recent cardiac surgery—then vancomycinshould be used instead of penicillin. This will cover most Staphylococcusepidermidis and methicillin resistant Staphylococcus aureus(MRSA).
  • Penicillin allergy—If the patient is allergicto penicillinthen a macrolide antibiotic, such as vancomycinor teicoplanin,should be used instead of penicillin.

Treatment in specific situations