Valvular heart disease

Rheumatic fever

Although it is less common than some decades ago, the combination of arthritis and a murmur in a febrile younger person should raise the possibility of rheumatic fever. Jones’ criteria have been revised several times to favour the decreasing prevalence in higher income countries (decreasing sensitivity and increasing specificity to avoid overtreatment.) A Brazilian study¹ found that (often due to late presentation) almost two thirds of cases of ARF would be missed if there was strict reliance on demonstrating a Group A streptococcal infection (AHA criteria.)

Jones criteria

Feature	Jones criteria
Occur in more than 55%
Carditis	Major
Arthritis	Major
Elevated ESR	Minor
ASOT > 320 IU/ml	Minor
Occur in 30 to 45%
Fever	Minor
Arthralgia	Minor
Chorea	Major
Occur in <3%
Subcutaneous nodules	Major
Erythema marginatum	Major
Increased PR on ECG	Minor

To make the diagnosis, you need two major criteria, or one major and two minor criteria, and in areas of low prevalence a positive ASOT or equivalent is required as well (AHA criteria.) In high prevalence areas such as most of public sector South African healthcare) it is reasonable to drop this requirement (WHO criteria. )

Management is with bedrest and analgesia (non-steroidal anti-inflammatory agent such as ibuprofen for arthritis.) Corticosteroids do not modify the natural history of the carditis.

Post streptococcal reactive arthritis.

This is an arthritis with positive streptococcal serology (ASOT or Anti-DNAase B) but it typically described as non-migratory, persisting for months, responding poorly to NSAIDs, and being more common in women. Before making this diagnosis, there should be echocardiographic evidence of no cardiac involvement, and obviously insufficient Jones criteria to make a diagnosis of ARF.

Sydenham’s chorea

This is quite rare in adults but in some series is found in 30% of cases of ARF, but is probably part of the spectrum of post-infectious autoimmune neurological disease which is uneasily linked by the same acronym (PANDA) to paediatric autoimmune neuropsychiatric disorder, where the focus is more on tics and behavioural disturbances. A South African study found a spectrum of movement and neuropsychiatric disorders evolving from an original presentation with chorea, and found that this condition is less benign than often thought.

Reasons for antibiotics in patients with rheumatic heart disease

Indication	Medication	Comment
Treatment of acute streptococcal pharyngitis	Benzathine penicillin 1.2 MU IM as a single dose	Single dose. ARR 1.67, NNT 53 from systematic review of old, poor quality trials.
Treatment of acute rheumatic fever	Benzathine penicillin 1.2 MU IM as a single dose	Conventionally given; weak evidence of efficacy
Prevention of further episodes of pharyngitis	Benzathine penicillin 1.2 MU IM	Every 3 to 4 weeks
Prevention of endocarditis in patients with established valvular heart disease	Amoxicillin 2 g orally one hour before procedure	Single dose; evidence of efficacy not clearly established
Empiric treatment of endocarditis before culture results available	Penicillin G 5 MU 4x/d IV plus gentamicin 120 mg (1.5mg/kg) 2x/d IV	Daily for 6 weeks, but modify according to blood culture results.

²

There are many recommendations on the duration of secondary prophylaxis, and they are seldom congruent with each other. As a general principle, patients with established carditis (i.e. a leaky valve) are thought to benefit from a longer duration of treatment.

No established carditis: until 21 years of age or for 10 years after last attack of ARF, whichever is longer (alternative: 18 years or 5 years after acute attack.)

Established carditis (usually mitral regurgitation): for at least 10 years after last attack of ARF, or till age 25 or 30 or 35. If carditis is “very severe”, then either till age 40 or life long.

This is clearly confusing. As a simplification, very few lower income country patients without carditis get a label of ARF, and most with carditis have severe disease. Many patients abscond from what is perceived as a very uncomfortable treatment, and setting patient achievable targets (e.g. age 25 or 30 years) may be pragmatic. An Australian study looking to improve adherence using a series of interventions addressed as quality improvement in clinics failed to demonstrate any benefit (80% injection adherence in 45% of patients….)³

Valvular heart disease.

Overview

Valvular problems in resource-constrained environments are usually due to rheumatic heart disease in childhood and adolescence. Management consists of correct identification of the lesion, and then timing the need for surgical intervention to balance risk against benefit. In general, lesions that lead to LV failure (aortic valve disease and mitral regurgitation) should be considered for surgery once symptomatic whereas mitral stenosis, which spares the LV, can have surgery delayed until symptoms aren’t controlled on furosemide in moderate doses. Even if asymptomatic, patients with LV damaging illnesses need regular echocardiographic review of LV function.

Cardiac examination is often considered very daunting by less experienced clinicians, and this state of affairs is hindered by a confusing array of eponymous physical signs, many of which contribute only marginally, if at all, to the conclusions from examination.

Positioning. It is quite remarkable how infrequently inexperienced auscultators move patients into the various positions that maximise murmur loudness. To dispense with procedures that more experienced clinicians find invaluable is a recipe for a disappointing auscultatory experience.

Aortic valve disease

The aortic valve can either be stenotic, leaking, or both. The systemic circulation is a high-pressure system, so disease of the aortic valve leads to increased LV work and eventually failure if the valve lesion is not corrected in time. Once the LV has failed, the risk of surgery rises very sharply and often becomes too dangerous to attempt.

Aortic stenosis:

• History of chest pain or fainting or shortness of breath.
• Low pulse pressure (small gap between systolic and diastolic pressures).
• Apex beat forceful (‘pressure loaded’, ‘sustained’) as it hypertrophies to eject stroke volume against increased resistance.
• Ejection systolic murmur at base that radiates into neck. The murmur may be quite soft, but is usually harsh, and also is occasionally heard quite well at the apex where it loses some of its lower frequency components and can sound rather like mitral regurgitation. (The Gallavardin phenomenon.)

Prognosis in aortic stenosis:

Patients tend to remain well for a long time with quite severe disease, but when they become ill, they may deteriorate rapidly. About half of symptomatic patients not operated on die within 2 years. Five year survival in asymptomatic AS is 65 to 95%, but once angina develops this drops to 50%. In patients with syncope, three year survival is 50%, and once heart failure develops, mean survival is only 1.5 years.

Surgical mortality is 1-8%, depending on the series recruitment (young patients with congenital or rheumatic disease versus elderly patients with associated coronary artery disease). Patients surviving to discharge have a 15% 5 yr mortality rate, i.e. 3% per year. However if there is a reduced ejection fraction, early (surgical) mortality can be as high as 25%.

Perspective – value of surgery in aortic stenosis

Selection bias is a problem with much of the literature in this field, but a widely quoted study⁴ perhaps puts it most dramatically: patients with severe symptomatic aortic stenosis who agreed to surgery had an 87% 3 year survival; those refusing surgery had a 21% 3 year survival. ARR = 66%, NNT = 2. In patients with LV dysfunction, the transvalvular gradient predicts outcome – a gradient of more than 40 mmHg is good, whereas a gradient of <30 mmHg is associated with a 50% 3 year mortality.⁵

Aortic regurgitation:

There is characteristically a large pulse pressure (diastolic pressure usually less than half the systolic pressure and often less than 50 mmHg). The apex is volume loaded, and there is a soft, rather inconspicuous early diastolic murmur at the aortic area and left sternal border, often heard best with patient leaning forward in inspiration.

Perspective – physical signs in aortic regurgitation.

A wide pulse pressure with a diastolic blood pressure of less than 50 mmHg (LR+ 19.3)⁶ and an early diastolic murmur make the diagnosis and are all you really need. An AR murmur to the left of the sternum should theoretically make you consider endocarditis with a ruptured cusp, syphilis, or aortic dissection (LR 8.2).⁷

Perseveration in the attempted elicitation of a veritable directory of eponymous signs is an example of confirmation bias – once you have elicited the wide pulse pressure in one way (checking the BP) repeating the test again and again does nothing for the posterior probability of the patient having aortic regurgitation, and in the interests of saving both time and sanity, a good case can be made for scrapping the lot. A review ⁸ looked at the importance that textbooks give these signs and found little correlation between their prominence and their reliability. Corrigan’s sign was labelled important in 10 of 11 textbooks and yet the literature reflects a sensitivity of 38 – 95%, and specificity (one study) of 16%.

For the record, and simply because it is fun to elicit, Duroziez’s sign is the combination of a systolic and diastolic murmur heard by pressing the stethoscope over the femoral artery. To make the diastolic component louder, press with a finger on the artery just distal to the stethoscope. This sign is perhaps still of value in patients with acute aortic regurgitation (usually secondary to endocarditis or dissection) in whom the murmur may be difficult to hear.

Indications for surgery in aortic valve disease.

Patients with aortic valve disease who are symptomatic warrant consideration for valve replacement surgery. Medical treatment doesn’t work once the patient is symptomatic and medication is not needed before then. In patients who need but cannot have surgery, judicious use of diuretics may help a bit, but don’t overdo them (postural hypotension). Do not use vasodilators unless in consultation with a specialist.

Mitral valve disease

Mitral regurgitation behaves like aortic valve disease – it causes LV damage if neglected. Mitral stenosis, on the other hand, protects the LV and rather leads to right sided damage with pulmonary hypertension as a late sequel.

Mitral regurgitation.

Unlike mitral stenosis, the first sound is a less important feature in the clinical diagnosis. One source ⁹ quotes it as being normal 75% of the time, loud in 12% and soft in 12% (in chronic MR). A loud S1 should cause one to think about mitral valve prolapse. A late systolic parasternal lift is supportive of MR (rather than pulmonary hypertension secondary to MS) as it is thought to be due to the LA increasing in size as systole continues, thus moving the heart up on an inflating cushion…

The presence of a characteristic murmur is strongly supportive of the diagnosis (LR 5.4) and a loud murmur (3/6 or more) has a LR+ for moderate/severe MR of 4.4¹⁰ (LR- i.e. murmur <3/6 was 0.2 for the presence of moderate or severe MR.) The presence of a third heart sound is also weakly predictive of moderate/severe MR (LR+ 1.8) but its absence does not preclude this (LR- 0.8).

Patients can initially be managed on furosemide, with or without the addition of a vasodilator (assuming no stenotic lesions!) Whether the addition of a vasodilator is really a good idea is somewhat moot, with some authors arguing that those who would benefit from a vasodilator would benefit more from valve replacement.

Valve repair or replacement should be considered once the patient is more than Class II disabled, and delay until LV function is too poor to tolerate surgery should be avoided. The mortality rate for patients not operated on but who have severe symptomatic mitral regurgitation is about 5% per year. Any increase in heart size on CXR should prompt referral for echocardiography so that LV dimensions can be measured and monitored. The echocardiographic criteria now commonly used as indications for surgery are an end-systolic LV diameter of more than 45 mm, or an LV EF of less than 60%.

Perspective – valve repair or replacement?

In predominantly US series with a low prevalence of rheumatic fever repair appears preferable to replacement¹¹ (post-operative ejection fraction fell by 10% if the valve was replaced, but was unchanged with repair if the chordate were preserved. Similarly, peri-operative mortality was 5 – 10% for replacement and only 2 – 4% for repair. Long term (5 – 10 year) survival was 80 – 94% for repair and 40 – 60% for replacement.

Perspective – NYHA grading of cardiac symptoms

The NYHA scale is widely used, but has been criticised for poor reproducibility. One study found a kappa of 0.24,¹² which is not very good. However, it does have the advantages of simplicity and familiarity, although every article seems to describe it a little differently. Here’s another version:

1 – no limitation of physical activity

2 – symptoms on walking

3 – symptoms on dressing

4 – symptoms at rest.

Prognosis in mitral regurgitation.

Older studies suggest a 5 year survival between 45 and 80%. With surgery, 5 year survival for rheumatic heart disease associated MR (as opposed to ischaemic) is about 75%, with rates varying considerably according to patient selection (young persons with good LV function and good compliance with warfarin will do better than older patients with poor LV function). The ‘ideal’ operation is valve repair but this is often not feasible with very distorted rheumatic valves and so valve replacement is usually performed.

Mitral stenosis.

The ‘tapping’ character of the apex in mitral stenosis is often easily discernable and is worth learning to elicit. A loud first heart sound is said to occur in 90% of patients. The opening snap occurs just after the second sound, in diastole. The murmur is also diastolic and is quite soft – listen at the apex with the bell of stethoscope in a patient lying on the left side.

Apparently stable patients with mitral stenosis may deteriorate quite abruptly. Common causes for this include:

• Pregnancy.
• Fever and tachycardia associated with chest infection.
• Anaemia.
• Thyrotoxicosis.
• Sudden onset atrial fibrillation.

Prognosis in mitral stenosis.

Reliable studies on the natural history of mitral stenosis in resource-constrained environments are difficult to find. Work from the pre-surgery era suggests ¹³ that the five year survival for NYHA class II was 80%, for class III 60%, and only 15% for those who were class IV disabled.

Various surgical techniques for dealing with mitral stenosis exist, ranging from closed valvotomy, to open commissurotomy to valve replacement. Early mortality is reported to be from 1-3%, and was zero in a US series of 120 patients undergoing open commissurotomy.¹⁴ The same series reported a 10 year actuarial survival of 95%, although mean follow-up was only 53 months. The re-stenosis rate was 7.5% overall, or 1.7% per year.

Indications for surgery in mitral stenosis.

Surgery is indicated once the patient is symptomatic, Class III or IV.

Frequency of follow-up in valvular heart disease

This frequency refers to follow-up by a medical officer and not by a cardiologist (all of whom are entirely capable of setting their own follow-up frequency!) The reason for this section is that some patients get ‘lost’ in the system and end up re-appearing when surgery is critically urgent or too late, whereas other remarkably stable patients seem to see a doctor monthly for no good reason other than that they are ‘cardiacs’.

Mitral stenosis:

In pure mitral stenosis, follow-up frequency can be determined primarily based on symptoms. Patients should be instructed to report to a doctor if they start to develop worsening effort tolerance while adherent to therapy. If they need more than 80 mg per day of furosemide to maintain effort tolerance, they should also ask to see a doctor and should be referred for assessment for surgery. A yearly visit to ensure that severe symptoms aren’t being overlooked in the particularly stoic is probably appropriate.

The rest:

• Patients with AS, AR, and MR should be reviewed more regularly by a doctor.
• If there is no clinical evidence of LV damage and the patient is clinically completely well, review after three months.
• If there are any symptoms, or any clinical evidence of LV dysfunction, the patient should be referred for baseline echocardiographic assessment of LV as a prelude to surgery. If echocardiography is unavailable except at the surgical referral site, refer for consideration for surgery anyway.
• If LV function at echocardiography is normal and surgery is not indicated (see above for indications for surgery in specific valve lesions), then review in three months.
• If LV function is abnormal, review monthly if surgery is not indicated yet, but the reason for deferring surgery should be very clearly understood and make sense to you. If you are not sure, then clarify by asking – it might be a misunderstanding.
• If surgery is not feasible (see above) then stabilise on medical therapy and review more regularly once symptoms worsen, as they inevitably will.

Summary: Patients with AS, AR, or MR in whom surgery is an option should be seen at least every three months if clinically and echocardiographically well, and more often if sick or with LV dysfunction. The recurrent question should be ‘does this patient need surgery now?’

Bioprostheses versus mechanical prostheses.

• In patients in whom adherence to warfarin therapy is problematic for a variety of possible reasons, alternatives are:
• To defer surgery, accepting that the complication rate from surgery plus poor adherence may be worse than the generally poor results associated with medical therapy of Class III or IV limited patients with almost any valve lesion.
• To accept suboptimal anticoagulation – e.g. with fixed dose warfarin.
• To insert a bioprosthetic valve, where the risk of thromboembolic complications after the first few months (sewing ring endothelialisation complete) is so low (about 1 – 2% per year) that anticoagulation is unjustified.
• Most clinicians have anecdotes of bioprosthetic valves that have lasted for decades against all odds, and of others that have degenerated within a couple of years or even less. A representative overall rate is about 2-3 % per year¹⁵ (aortic valves).

Perspective – warfarin adherence.

Many clinicians automatically assume that INR fluctuations are because of ‘patient failure’. In one study¹⁶ of 347 patients and 4305 INRs, 23% were outside the ‘desirable’ range (1.8-3.4), but in only 35.7% was this because of non-adherence. In a similar proportion (37.6%) the cause was unclear, and in just more than one quarter, the cause was either medical system failure (9.1%) or change in the clinical status of the patient (17.6%.)

Particularly instructive is a breakdown of those definitely ascribed to non-adherence: half were due to a change in dietary intake of vitamin K, one quarter to patients not understanding new dosage instructions (arguably actually a medical system failure) and one quarter (only 9% of the total 1002 out of range INRs) due to ‘traditional’ poor adherence issues such as forgetting doses, forgetting clinic visits, forgetting to pick up repeat scripts or ethanol use. Put differently, 91% may have been due to problems not entirely under the patient’s control…

Perspective – warfarin and herbal medication.

Local data on this are scanty. In a postal survey conducted in the UK ¹⁷ (only 54.2% response rate), 19.2% of patients on warfarin were taking one or other herbal medication and 8.8% were taking medication known to interact with warfarin. More than 92% of responders had never discussed the use of herbal medication with their doctors.

Many of the commercial herbal medications (garlic, ginseng, feverfew and ginger) are thought to lead to a prolonging of the INR. St John’s wort may reduce the INR. Where patients’ diet is seasonally dependent, availability of Vitamin K containing vegetables may vary considerably, and this should be taken into account when trying to determine the cause of puzzling fluctuations in the INR. It also does no harm to ask about herbal medications – patients are often surprisingly candid once they trust you.

Pre-operative work-up of patients going for cardiac surgery

Delays in busy surgical units can be due to finding patients with active rheumatic carditis, unsuspected endocarditis, unsuspected tuberculosis, severe dental caries, or who are vigorously anticoagulated. Check the following before sending:

• CRP, ASOT and HIV status.
• Dental attention.
• CXR and clinical screen for pulmonary tuberculosis.
• It is also helpful to send copies of the local clinical notes with the patient.
• Patients with ischaemic heart disease usually warrant consideration for stress testing prior to referral, and it is advisable to send a copy of the tracing with the patient.
• Stop warfarin about two days before transfer if cardiac catheterisation is indicated.

Post-op follow-up

Documentation

Documentation of what was done, why it was done and when it was done can be extremely useful at a later stage. Always make notes to this effect on any patient returning from surgery.

Physical findings

Listen to the heart as well as checking pulse (sinus?) and BP on returning from surgery. Write down what you find. If the patient becomes sick and there is a new murmur on auscultation at a later stage, this baseline information can be valuable.

Anti-coagulation

All new valves for three months. With bioprosthetic (tissue) valves you can then stop unless the patient is in atrial fibrillation or has another reason to remain on anticoagulation. Aortic metallic valves have less of a thromboembolic risk than mitral, but valves in both positions should be managed with INR at least 2 and preferably closer to 3.

Frequency of follow-up

Initially weekly until the INR is clearly under control. There is often considerable movement and need for adjustment in the first few weeks as patients reduce doses of analgesics and settle into home routines. Once stable, INR monthly is desirable, but fixed dose management may be all that is possible.

Suspected valve dysfunction

Consider that there may be a problem with the prosthetic valve if:

• A patient develops any new cardiorespiratory symptoms not immediately attributable to another cause (e.g. a pneumonia.)
• A patient develops worsening or new cardiac failure signs.
• New anaemia develops (haemolysis, often from a paravalvular leak)
• New murmurs develop (mechanical valves are a little ‘stenotic’ so it is normal to hear an ejection systolic murmur with an aortic valve and a mid-diastolic murmur with a mitral valve replacement.)
• A click which was previously easily heard becomes difficult to hear or inaudible.
• New fever or malaise (SBE)
• Suddenly rising INR in previously stable patient (often a sign of cardiac failure with liver congestion, so examine the patient!)

Dental work after valve replacement

Patients needing scaling or tooth extraction (not just a filling) or at a very small but increased risk of bacteraemia and subsequent endocarditis, so many authorities recommend manage with intravenous ampicillin 2g immediately before the procedure and again 6 hours later, combined with gentamicin 120 mg once at the same time as the first dose of ampicillin. The evidence that this is efficacious is tenuous. Stop warfarin a couple of days beforehand so that the INR can settle to about 1.4 or less, then re-start it immediately after the procedure. Local bleeding can often be controlled with tranexamic acid mouth washes, which may allow more adventurous practitioners to proceed at relatively higher INRs, but in general err on the side of caution.

Drug interactions

There are a considerable number of relatively unfamiliar drug interactions that may occur in patients with HIV. Key ones to consider are warfarin and fluconazole, co-trimoxazole, efavirenz and the protease inhibitors, but many others are described. If in doubt, look it up!

Functional murmurs.

These are early systolic murmurs usually at the LSB, and < 2/6 in intensity. There is no other evidence of cardiac pathology (no diastolic murmur, gallop, raised JVP, displaced apex) and the ECG and CXR are normal. They are common in pregnancy.

Supraclavicular venous bruits – change with arm or head re-positioning.

Endocarditis

Endocarditis natural history

The ‘classic’ form of endocarditis seen in resource-constrained countries is becoming less common in more affluent countries, where prosthetic valve endocarditis, endocarditis in intravenous drug abusers, and nosocomial endocarditis are becoming more common. A literature review found that the in-hospital mortality was still about 16% (range 11-26%).¹⁸

Endocarditis prophylaxis

Indications: Prophylaxis may be indicated for dental procedures involving gingival manipulation but is no longer recommended for genitourinary or gastrointestinal procedures. Patients at highest risk of complications if they were to develop endocarditis are considered to be those with prosthetic valves, previous endocarditis, unrepaired congenital heart disease, or repaired disease with residual shunts. It may be reasonable to consider prophylaxis in such patients:

Amoxicillin 2g (4 capsules of 500 mg) PO 1 hour before the procedure. (if nil per mouth, give ampicillin 2g IV.)

If allergic to penicillin, clindamycin 600 mg PO 1 hour before the procedure (same dose IV if nil per mouth.)

Perspective – should we be using endocarditis prophylaxis at all?

A 2008 UK review ¹⁹ raises a concern that while the evidence of benefit from infective endocarditis prophylaxis is not firmly established, there is clearer data on the harms of antibiotic use. It is argued that these risks (e.g. the low risk of anaphylaxis) outweigh any currently measured benefits. The point is also made that bacteraemic events occur very frequently (e.g. from tooth brushing) and hence ‘once off’ prophylaxis is unlikely to influence life-time risk. The arguments are cogent, and a case for withholding antibiotics because they do more harm than good is entirely defensible but currently still contrary to much medical tradition.

As a transitional recommendation, until either the evidence of benefit becomes clearer, or the enthusiasm for use wanes, it may be reasonable to continue use in ‘high risk’ patients.

Natural history

The ‘classic’ form of endocarditis seen in resource-constrained countries is becoming less common in more affluent countries, where prosthetic valve endocarditis, endocarditis in intravenous drug abusers, and nosocomial endocarditis are becoming more common. A literature review found that the in-hospital mortality was still about 16% (range 11-26%^[32]).

Perspective – organisms in endocarditis

The review quoted above found that in a series of original valve endocarditis patients, the most frequent organism was staphylococci (44%), followed by streptococci (31%), and then much less commonly enterococci (9%), HACEK group organisms (4%) and fungi (1%.)

The HACEK organisms are a group of gram negatives: Haemophilus, Actinobacillus, Cardiobacterium, Eikenella and Kingella.) Other organisms involved in so-called culture negative endocarditis include Bartonella, Legionella, Brucella and Chlamydia. In infections due to any of the stranger organisms, it is important to discuss therapy with a clinical microbiologist of infectious disease expert.

Diagnostic process in suspected endocarditis.

The diagnosis of infective endocarditis is made or dismissed on clinical grounds in the majority of patients. It is easy enough if a fresh fat vegetation is seen on echocardiography in a febrile patient with haematuria and a murmur. In the majority of patients, things are not that simple.

• Examine the patient meticulously on a daily or alternate day basis.
• All peripheral pulses still present and equal?
• Any new splinters? (Write down findings immediately – it is easy to forget on which finger you saw that now vanished little mark…)
• Conjunctival haemorrhages? Fundi?
• New murmur or changing murmur?
• Haematuria on dip sticks?
• New arthritis? New neurological signs or symptoms?
• New splenomegaly.
• Fever – check the temperature yourself in any doubt.
• Rising CRP or ESR.
• Leukocytosis.
• New increasing PR interval or bundle branch block on ECG (aortic root abscess?)

Duke’s criteria.

Duke proposed a list of criteria to be fulfilled in order to make a diagnosis of endocarditis. Essentially, this is the familiar process of applying a series of likelihood ratios one after the other in order to revise a probability. The criteria are:

Major criteria

• Positive blood culture
• Positive echocardiogram
• New regurgitant murmur

Minor criteria

• Predisposing heart condition
• Temperature > 38C
• Vascular phenomena – e.g. splinters, mycotic aneurysm, intracranial bleed, Janeway lesions
• Immunological phenomena – glomerulonephritis, Roth spots, Osler’s nodes
• Abnormal but non-‘diagnostic’ echocardiogram

Ways of achieving ‘probable’ diagnosis:

Two major OR

One major and three minor OR

Five minor.

Perspective – value of individual Duke criteria

It is fairly clear that major criteria that are positive have good diagnostic value; what is less well emphasized is that the very rare minor criteria such as Roth spots, Osler’s nodes and Janeway lesions are well know but very seldom seen (sensitivity thus of the order of 3 to 5%) they are useful (very specific) when present:

This emphasizes the importance of careful, repeated physical examination, multiple blood cultures, and the need to remain wary of making the diagnosis based predominantly on minor criteria.

Althouth Osler’s nodes and Janeway lesions are usually considered vasculitic/immunologic they are probably embolic, with what weak biopsy evidence there is favouring microabscess formation in aetiology²⁰

Perspective – echocardiography in suspected endocarditis

Transoesophageal echocardiography is best but difficult to access. Conventional transthoracic echo is unsuccessful due to poor visualisation in about 20% of people, mainly because of COPD or obesity. What this means is that the test is good for ruling in but bad at ruling out the diagnosis. (See table.) Results for transoesophageal echo are superior, with some authors reporting sensitivity and specificity of 85% and 97% respectively, although there were often problems with study design.

Test	Sensitivity	Specificity	LR+	LR-
Transthoracic echocardiography	0.60	0.98	30.0	0.4
Blood culture	95%
Duke’s criteria	0.76	0.97	28.4 (7.2-112.2)	0.24 (0.16 –0.37)

²¹,²²

(In the study of Duke’s criteria referenced above, the sensitivity and specificity figures were arrived at by classifying ‘possible’ cases as negative; if considered positive, sensitivity was higher (100%) but specificity fell to 88%. ‘Possible’ cases were those which seemed like endocarditis but didn’t fulfill the criteria above, but where the diagnosis could not be firmly excluded, e.g. by autopsy, surgery, a firm alternative explanation, or resolution of the syndrome with 4 or fewer days of antibiotics.)

However, in a large retrospective cohort study using tighter criteria for a ‘negative’ echocardiogram (adequate ultrasound quality, no valve stenosis, trivial regurgitation, no or small pericardial effusion, normal anatomy, no vegetations on valves, and no ‘hardware’, including no catheters) sensitivity of transthoracic echo was 98% and negative predictive value 97% (versus 43% and 87% respectively using the traditional ‘no vegetation seen’ approach.) ²³

Diagnostic alternatives

Simple things need to be considered: a person with mitral regurgitation who comes in with fever, haematuria and dysuria may well simply have a UTI. Acute viral illnesses (flu!) can cause a lot of concern, but usually settle quite quickly. An HIV seroconversion illness is another diagnostic possibility. One of the major sources of confusion in resource-constrained environments is megaloblastic anaemias. Patients present with fever, anaemia, a systolic flow murmur (due to the anaemia) and sometimes even fundal haemorrhages. Check the MCV, particularly if the Hb is less than 7g%.

Therapy – endocarditis.

Always try to get three blood cultures (at least) out of the patient before starting therapy. Infective endocarditis is usually a sub-acute process, so you can take a day or so to take the cultures. Culture negative endocarditis has a worse prognosis that that where the organism is identified – you can’t target therapy. One observational cohort study ²⁴ found that the mortality rate when clinicians had waited 72 hours before starting therapy (and taken blood cultures during that time) was 18% compared to 30% (however p=0.18 for the difference) when therapy was started early, but the bias may have been to start earlier in sicker patients.

Empiric therapy before culture results available:

• Penicillin G: 5 million units 6 hourly for 4 weeks PLUS
• Gentamicin: 1.5 mg/kg every 12 hours – e.g. 80 mg 12 hourly. This is a synergistic dose. Give it for the first two weeks.

Remember that high dose penicillin delivers a significant sodium load, so you may need to increase diuretic doses.

Once you have the organism, continue with the above if it is a sensitive streptococcus, and swap to cloxacillin 3g 6 hourly plus genta 1.5 mg/kg every 12 hours if it is a staphylococcus. In patients who are allergic to penicillin, vancomycin 1g 12 hourly is a reasonable alternative.

Duration: In all cases therapy should be for 4 weeks. In exceptional cases, where there is bed pressure and you have identified a viridans group streptococcus which is penicillin sensitive and has responded well clinically, there is some evidence to support swapping to amoxicillin 1g 12 hourly orally after the end of the second week, and continuing with this for a further 4 weeks, but in general this is discouraged.

Consult a microbiologist or refer the patient if an usual organism is cultured.

Patients with prosthetic valve endocarditis should be referred early to a cardiac centre, but if empiric therapy is needed, consideration can be given to a combination of vancomycin 1g 12 hourly IV (30 mg/kg/d) plus rifampicin 450 mg 2x/d PO (15 mg/kg/d) plus gentamicin 80 mg 12 hrly IV (3 mg/kg/d.) The gentamicin is usually only given for the first two weeks.

---

1. Pereira BA, da Silva NA, Andrade LE, et al. Jones criteria and underdiagnosis of rheumatic fever. Indian J Pediatr. 2007;74(2):117-121 ↩

2. Robertson KA, Volmink JA, Mayosi BM. Antibiotics for primary prevention of acute rheumatic fever: a meta-analysis. BMC Cardiovascular Disorders2005, 5:11 ↩

3. Ralph AP, de Dassel JL, Kirby A, et al. Improving Delivery of Secondary Prophylaxis for Rheumatic Heart Disease in a High‐Burden Setting: Outcome of a Stepped‐Wedge, Community, Randomized Trial. https://doi.org/10.1161/JAHA.118.009308Journal of the American Heart Association. 2018;7:e009308 ↩

4. Schwarz F, Baumann P, Manthey J, et al. The effect of aortic valve replacement on survival. Circulation. 1982;66:1105-10 ↩

5. Carabello BA. Aortic stenosis. N Engl J Med 2002;346:677-82 ↩

6. McGee S. Evidence-based physical diagnosis. WB Saunders 2001. p539 ↩

7. Harvey WP, Corrado MA, Perloff JK. “Right-sided” murmurs of aortic insufficiency.  Am J Med Sci 1963; 245:533-43. Quoted by McGee S in Evidence-based physical diagnosis. ↩

8. Babu An, Kymes SM, Carpenter Fryer SM. Eponyms and the diagnosis of aortic regurgitation: what says the evidence? Ann Intern Med 2003;138:736-42 ↩

9. McGee S. Evidence based physical diagnosis. WB Saunders 2001 p547 ↩

10. Desjardins VA. Intensity of murmurs correlates with severity of valvular regurgitation. Am J Med 1996;100:149-56 ↩

11. Otto CM. Evaluation and management of chronic mitral regurgitation. N Engl J Med. 2001;345:740-6 ↩

12. Visser MC, Koudstaal PJ, van Latum JC, et al. Interobserver variation in the application of 2 disability scales in heart patients. Ned Tijdschr Geneeskd. 1992;136:831-4 ↩

13. Braunwald E. A Textbook of Cardiovascular Medicine. WB Saunders 1992. p1015 ↩

14. Cohn LH, Allred EN, Cohn LA, et al. Long-term results of open mitral valve reconstruction for mitral stenosis. Am J Cardiol. 1985;55:731-4 ↩

15. Feyrer R, Lott S, Englmann P, et al. Results of a 1 to 31 years follow-up after aortic valve replacement in 2327 patients. J Cardiovasc Surg. 1996;37:17-22 ↩

16. Waterman AD, Milligan PE, Bayer L, et al. Effect of warfarin nonadherence on control of the International Normalised Ratio. Am J Health-Syst Pharm. 2004;61:258-64 ↩

17. Smith L, Ernst E, Ewings P, et al. Co-ingestion of herbal medicines and warfarin. Br J Gen Practice. 2004;54:439-41 ↩

18. Moreillon P, Que Y. Infective endocarditis. Lancet 2004;363:139-148 ↩

19. NICE Clinical Guideline 64: Prophylaxis against infective endocarditis. www.nice.org.uk/CG064. ↩

20. Alpert JS. Osler’s nodes and Janeway lesions are not the result of small-vessel vasculitis. Am J Med. 2013 Oct;126(10):843-4. doi: 10.1016/j.amjmed.2013.04.002. PMID: 24054951 ↩

21. Mylonakis E, Calderwood SB. Infective endocarditis in adults. N Engl J Med 2001;345:1318-30 ↩

22. Cecchi E, Parrini I, Chinaglia A, et al. New diagnostic criteria for infective endocarditis. A study of sensitivity and specificity. European Heart J 1997; 18: 1149-1156 ↩

23. Sivak JA, Vora AN, Navar AM, et al. An Approach to Improve the Negative Predictive Value and Clinical Utility of Transthoracic Echocardiography in Suspected Native Valve Infective Endocarditis. J Am Soc Echocardiogr. 2016 Apr;29(4):315-22. doi: 10.1016/j.echo.2015.12.009. Epub 2016 Feb 3. PMID: 26850679; PMCID: PMC6052444 ↩

24. Koegelenberg CF, Doubell AF, Orth H, et al. Infective endocarditis: improving the diagnostic yield. Cardiovasc J S Afr. 2004;15:14-20 ↩

Was this article helpful?

0 out Of 5 Stars

5 Stars		0%
4 Stars		0%
3 Stars		0%
2 Stars		0%
1 Stars		0%

5 Submit

Send Feedback

How can we improve this article?

Please submit the reason for your vote so that we can improve the article.

Submit