Which patients should undergo preoperative non-invasive investigations or coronary angiography?

Non-invasive testing refers to investigations other than angiography such as dipyridamole thallium scanning or dobutamine stress echocardiography. The literature on this question is overwhelming. It is best approached by nine simple steps. These are based on the recommendations of the joint consensus conference of the American College of Cardiology and the American Heart Association.1 Clinical predictors, functional capacity and magnitude of surgical risk can be assessed from Tables 101.3, 101.4 and 101.5 in the next question.

Step 1 What is the urgency of surgery?
If absolute emergency proceed to surgery, otherwise proceed to step 2.

Step 2 Has the patient undergone coronary revascularisation in the last five years?
If so and symptoms are stable, proceed to surgery. If not, or symptoms are unstable go to step 3.

Step 3 Has there been a coronary evaluation in the past two years?
If so and there are no changes or new symptoms proceed to surgery.
If not, or there have been changes go to step 4.

Step 4 Is there an unstable coronary syndrome or major clinical predictor of risk?
If so proceed direct to angiography. If not go to step 5.

Step 5 Are there intermediate clinical predictors of risk?
If so go to step 6. If not go to step 7.

Step 6 What is the functional capacity and magnitude of surgical risk?
If there are intermediate clinical predictors, then order noninvasive investigations if there is either poor function or high surgical risk. Otherwise go to surgery.

Step 7 Are there minor clinical predictors?
If so go to step 8. If not proceed to surgery.

Step 8 What is the functional capacity and magnitude of surgical risk?
If there are minor clinical predictors, then order non-invasive investigations if there are both poor function and high surgical risk.

Step 9
All patients have now been assigned to surgery, angiography or non-invasive testing. The results of non-invasive tests must incorporate both the absolute result (positive or negative) and quantification of the result (e.g. magnitude and regional location of ischaemic area). These results will determine which patients should proceed to angiography. Significant abnormalities require further assessment by angiography. Minor and intermediate abnormalities only require further assessment in the presence of low functional capacity or major surgical risk. It should be noted that, at least in high-risk patients undergoing vascular surgery, beta blockade is the only medical intervention proven to have major impact on outcome.

Which patients should receive antibiotic prophylaxis for endocarditis, and which procedures should be covered in this way?

There is little firm scientific evidence for present advice on antibiotic prophylaxis for endocarditis, mainly because of the rarity of the disease. Only 10% of cases are related to bacteraemia caused by invasive procedures. Prevention of endocarditis in patients with abnormal heart valves can be achieved by many general measures, for example, regular dental care. The convention for the use of antibiotics in the prevention of endocarditis derives from animal models and clinical experience. Although dental extraction results in a bacteraemia of about 100cfu/mL, no obvious relationship has been found between the number of circulating bacteria and the likelihood of developing endocarditis.

In man, case-control studies suggest 17% of cases might be prevented if prophylaxis is given for all procedures in patients with abnormal valves.1 Individual cases of endocarditis following dental or urological procedures have been reported but the risk of developing endocarditis must be very low. Underlying cardiac abnormalities greatly increase the risk of endocarditis, e.g. patent ductus arteriosus, prosthetic valves, hypertrophic cardiomyopathy, aortic valve disease or previous endocarditis.
Mitral valve prolapse is common but merits antibiotic prophylaxis if it causes a murmur.

Procedures causing gingival bleeding should be covered by prophylaxis as should tonsillectomy, adenoidectomy and dental work. Other procedures in which prophylaxis should be used include oesophageal dilatation or surgery or endoscopic laser procedures, sclerosis of oesophageal varices, abdominal surgery, instrumentation of ureter or kidney, surgery of prostate or urinary tract. Flexible bronchoscopy with biopsy, cardiac catheterisation, endoscopy with biopsy, liver biopsy, endotracheal intubation and
urethral catheterisation in the absence of infection do not need prophylaxis. Patients having colonoscopy or sigmoidoscopy probably do not require prophylaxis unless there is a prosthetic valve or previous endocarditis or unless biopsy is likely to be performed. Recommendations for prophylaxis in patients undergoing obstetric or gynaecological procedures are required for patients with prosthetic valves, or who have previously had endocarditis.

Recommendations for prophylaxis vary between countries. Dental (causing gingival bleeding), oropharyngeal, gastrointestinal and urological procedures are usually considered a risk.2 The use of antibiotic prophylaxis is routine during cardiac surgery, flucloxacillin, plus an aminoglycoside, or a cephalosporin being common choices.

What percentage of blood cultures will be positive in endocarditis?

The great majority of patients with endocarditis have positive blood cultures within a few days of incubation and only a few cases will become positive on further incubation for 1–2 weeks. The proportion of culture-negative cases depends on the volume of blood and method of culture but a common estimate is 5% with a range from 2.5% to 31%.1 Most cases of culture-negative endocarditis are related to use of antibiotics within the preceding two weeks and probably represent infections with staphylococci, streptococci or enterococci. If antibiotics have been given, withdrawal of treatment for four days and serial blood cultures will usually demonstrate the pathogen.

A number of organisms may grow only if incubated under the correct conditions. Nutritionally-deficient streptococci may fail to grow in ordinary media and yet are part of the normal mouth flora and can cause endocarditis.2 The HACEK organisms are slow
growing and easily missed. Coxiella burnetti, Chlamydia spp. and Mycoplasma spp. are rare causes of endocarditis and are difficult to grow, diagnosis requiring biopsy or serology. Bartonella spp. are now known to cause endocarditis in homeless patients and diagnosis
is difficult by conventional methods.

Three sets of blood cultures will demonstrate at least 95% of culturable organisms causing endocarditis. After four negative cultures there is only a 1% chance of an organism being identified by later culture.4 Contamination as the result of poor collection technique makes interpretation difficult and is a greater risk when repeated sets of culture are collected.

What is the morbidity and mortality of endocarditis with modern day management (and how many relapse)?

Despite progress in management, morbidity and mortality remain major problems for the patient with endocarditis, both during the acute phase and as the result of long term complications after a bacteriological cure. Improvements in microbiological diagnosis, types of antibiotic treatment and timing of surgical intervention have improved the outlook for some patients but the impact has been minor with some of the more invasive pathogens. The infection can relapse and vegetations can be reinfected. Healed vegetations may leave valvular function so compromised that surgery is required.

In 140 patients with acute infective endocarditis, 48 (34%) required valve replacement during treatment.1 Heart failure occurred in 46 patients. During the active disease, 22 patients (16%) died. Medical treatment alone cured 80 patients. Relapse occurred in 3 (2.7%) of 112 patients all within one month of discharge. Recurrence was observed in 5 (4%) patients between 4 months and 15 years after the first episode. In the follow up period, another 16 patients died of cardiac causes, most within five years. Of 34 patients with late prosthetic valve endocarditis, 27 (79%) survived their hospital admission but 11 had further surgery during the next five years, usually following cardiac failure.2 In another study, 91 (70%) of 130 patients survived hospitalisation for native valve endocarditis and 17 of 60 initially treated medically required surgery during a mean 9 year follow up.3 During follow up, 29 (22%) patients died, 13 from cardiac causes.

What are the indications for surgical management of endocarditis?

The indications for surgical management of endocarditis fall into six categories.

1. Congestive heart failure
   Patients with moderate-to-severe heart failure require urgent surgical intervention. With mitral regurgitation, afterload reduction and diuretic therapy can improve symptoms and may make it possible to postpone surgical repair until a full course of antibiotic therapy has been completed. In contrast, acute aortic regurgitation progresses rapidly despite an initial favourable response to medical therapy, and early surgical intervention is imperative.
2. Persistent sepsis
   This is defined as failure to achieve bloodstream sterility after 3–5 days of appropriate antibiotic therapy or a lack of clinical improvement after one week.
3. Recognised virulence of the infecting organism
   
   • With native valve endocarditis, streptococcal infections can be cured with medical therapy in 90%. However, S. aureus and gram negative bacteria are more aggressive, requiring transoesophageal echocardiography to rule out deep tissue invasion or subtle valvular dysfunction. Fungal infections invariably require surgical intervention
   • With prosthetic valve endocarditis, streptococcal tissue valve infections involving only the leaflets can be cleared in 80% with antibiotic therapy alone; however, mechanical or tissue valve infections involving the sewing ring generally require valve replacement. If echocardiography demonstrates a perivalvular leak, annular extension, or a large vegetation, early operation is necessary
4. Extravalvular extension
   Annular abscesses are more common with aortic (25-50%) than mitral (1-5%) infections; in either case, surgical intervention is preferred (survival: 25% medical, 60-80% surgical). Conduction disturbances are a typical manifestation.
5. Peripheral embolisation
   This is common (30-40%), but the incidence falls dramatically following initiation of antibiotic therapy. Medical therapy is appropriate for asymptomatic aortic or small vegetations. Surgical therapy is indicated for recurrent or multiple embolisation, large mobile mitral vegetations or vegetations that increase in size despite appropriate medical therapy.
6. Cerebral embolisation
   Operation within 24 hours of an infarct carries a 50% exacerbation and 67% mortality rate, but the risk falls after two weeks (exacerbation <10%,>

How should the anticoagulation of a patient with a mechanical heart valve be managed for elective surgery?

Mechanical heart valves are associated with an annual risk of arterial thromboembolism of <8%. Although warfarin greatly reduces the risk, it is at the expense of an INR-related risk of serious haemorrhage. This constitutes an unacceptable risk for patients undergoing major surgery, and it is necessary to temporarily institute alternative anticoagulant measures.

The anticoagulant effect of oral warfarin is prolonged (half life 36 hours) and it can take 3–5 days for a therapeutic INR to fall to less than 1.5. It is also dependent on the half life of the vitamin K dependent clotting factors (particularly factors X and II, with half lives of 36 and 72 hours respectively). The surgical procedure must therefore be planned with this in mind. A “safe” INR depends on the surgery being undertaken. An INR <1.5 is usually suitable, although this should be <1.2 for neurosurgical and
ophthalmic procedures.

Four days prior to surgery warfarin should be stopped. Once the INR falls below a therapeutic level heparin should be started. Unfractionated heparin (UFH) should be administered as an intravenous infusion. It has a short lasting effect (half life 2 to 4 hours) and is monitored using daily measurements of the APTT ratio (aim for APTT 1.5–2.5 times greater than control APTT). Alternatively, a weight-adjusted dose of low molecular weight heparin (LMWH) is given subcutaneously once daily with
predictable anticoagulant effect, although data are limited. The night prior to surgery the INR should be checked and if it is inappropriately high then surgery should be delayed. If surgery cannot be delayed, the effect of warfarin can be reversed by fresh frozen plasma (2–4 units) or a small dose of intravenous vitamin K (0.5–2mg). Six hours prior to surgery heparin should be stopped to allow the APTT to fall to normal.

Recommencing intravenous heparin in the immediate postoperative period may increase the risk of haemorrhage to greater levels than the risk of thromboembolism with no anticoagulation. Heparin is usually restarted 12–24 hours after surgery, depending on the type of surgery and the cardiac reason for warfarin. Each case must be considered individually. Warfarin should be restarted as soon as the patient is able to tolerate oral medication. Prophylactic heparin should be stopped once an INR greater than 2.0 is established.

A patient is on life-long warfarin and wishes to become pregnant. How should she be managed?

All anticoagulant options during pregnancy are associated with potential risks to the mother and fetus. Any woman on warfarin who wishes to become pregnant should ideally be seen for prepregnancy counselling and should be involved in the anticoagulation decision as much as possible. Potential risks to the fetus need to be balanced against the increased maternal thrombotic risk during pregnancy. Anticoagulation for mechanical heart valves in pregnancy remains an area of some controversy.

The use of warfarin during pregnancy is associated with a low risk of maternal complications1 but it readily crosses the placenta and embryopathy can follow exposure between 6–12 weeks’ gestation, the true incidence of which is unknown. A single study has reported that a maternal warfarin dose 5mg is without this embryopathy risk.2 As pregnancy progresses, the immature vitamin K metabolism of the fetus can result in intracranial haemorrhage even when the maternal INR is well controlled. In addition, a direct CNS effect of warfarin has been described, resulting in structural abnormalities. Conversion to heparin in the final few weeks of pregnancy is recommended to prevent the delivery of, what is in effect, an anticoagulated fetus.

In contrast, unfractionated heparin (UFH) is free from direct fetal harm but it has varied pharmacokinetic and anticoagulant effects and adequate maternal anticoagulation can be difficult to achieve. The use of UFH in women with mechanical valve replacements during pregnancy has been associated with increased maternal thrombosis and bleeding. Studies have been criticised for the use of inadequate heparin dosing and/or inadequate therapeutic ranges although a recent prospective study which used heparin in the first trimester and in the final weeks of pregnancy reported fatal valve thromboses despite adequate anticoagulation. Long term heparin use risks osteoporosis and heparin-induced thrombocytopenia (HIT). Intensive monitoring is required in pregnancy and the use of anti-Xa assays may be necessary.

Low molecular weight heparins (LMWH) have a more reliable anticoagulant effect.6 The dose is adjusted according to anti-Xa levels. Use in pregnancy is mainly for thromboprophylaxis rather than full anticoagulation but experience is increasing. Indeed, case reports are starting to emerge where LMWH has been used for mechanical valve replacements. Compared with UFH the risk of HIT and osteoporosis are reduced6 and these heparins may hold the future for anticoagulation in pregnancy.

Management
Women who do not wish to continue warfarin throughout pregnancy can be reassured that conceiving on warfarin appears safe but conversion to heparin, to avoid the risk of embryopathy, needs to be carried out by 6 weeks. Breast-feeding on either warfarin or heparin is safe. Possible regimes include:

• Warfarin throughout pregnancy until near term and then conversion to unfractionated heparin.
• Unfractionated heparin for the first trimester. Warfarin until near term and then resumption of heparin.

Which cardiac patients should never get pregnant? Which cardiac patients should undergo elective Caesarean section?

Which women should never get pregnant?

1. Those with significant pulmonary hypertension (pulmonary vascular resistance >2/3 of systemic), especially cyanotic patients and those with Eisenmenger reaction (maternal mortality ~50%) and those with residual pulmonary hypertension after e.g. VSD closure. NB: Even women with modest pulmonary vascular disease ~1/2 systemic are at risk of death.
2. Those with grade 4 systemic ventricular function (EF <20%).>

Which women should not get pregnant until operated upon?

1. Marfan’s syndrome patients with aortic aneurysm/dilated aortic root.
2. Those with severe left sided obstructive lesions (AS, MS, coarctation).

Which women should undergo elective Caesarean section?

1. 1 Those with independent obstetric indications.
2. 2 Caesarean section should be strongly considered for the following women:
   • Those with mechanical valves, especially tilting disc in the mitral position. The key here is to leave the mother off warfarin for the minimum time possible. An elective section is performed at 38 weeks’ gestation, replacing the warfarin with unfractionated heparin for the minimum time possible
   • Severe aortic or mitral stenosis.

If the mother’s life is at risk, section followed by valve replacement may be necessary.

Controversy remains over whether the following patients should undergo elective Caesarean section:

1. Cyanotic congenital heart disease with impaired fetal growth. Section may help to avoid further fetal hypoxaemia, but at the expense of excessive maternal haemorrhage to which cyanotic patients are prone.
2. Pulmonary hypertension. See comments above.
   

A balance has to be made between a spontaneous vaginal delivery with the mother in the lateral decubitus position to attenuate haemodynamic fluctuations, forceps assistance and the smaller volume of blood lost during this type of delivery, and the controlled timing of an elective section. Probably more important than the route of delivery is peri-partum planning and teamwork: delivery must be planned in advance, and the patient intensively monitored, kept well hydrated and not allowed to drop her systemic vascular resistance. Consultant obstetric and anaesthetic staff experienced in these conditions should be present, and the cardiologist readily available.

How do I manage the pregnant woman with valve disease?

Native or tissue valves
In general, regurgitant lesions are well tolerated during pregnancy, whereas left sided stenotic lesions are not (increased circulating volume and cardiac output lead to a rise in left atrial pressure). Tissue valves can deteriorate rapidly during pregnancy. Management of patients with significant mitral and aortic stenosis

1. Bedrest:
   
   • Reduced heart rate allows time for LV filling and ejection
   • Reduced venous return due to IVC compression by the uterus reduces LA pressure (also increases risk of thrombosis: patients must be heparinised).
2. Dyspnoea and angina: slow the heart rate with beta blockers or digoxin. Nitrates may be useful, but should be used with caution in those with aortic stenosis.
3. Intractable pulmonary oedema:
   
   • Balloon valvotomy
   • Closed mitral valvotomy (advantage as no cardiopulmonary bypass, but few surgeons nowadays have experience)
   • If valvotomy not possible, then deliver fetus by Caesarean section followed by cardiopulmonary bypass and valve replacement.

Mechanical valves
Anticoagulation is the issue here: in particular, the risk of warfarin embryopathy vs risk of valve thrombosis.

The choice lies between:

1. Warfarin throughout pregnancy, stopping it for a minimal length of time for delivery
2. Convert to heparin during the first trimester with hospital admission and meticulous control of APTT. Return to warfarin for the second trimester and reinstate heparin at ~34/40.

Note:

1. Mitral tilting disc prostheses at particular risk: fatal thrombotic occlusion of these valves in pregnant women described despite well-controlled heparin anticoagulation
2. Risk of significant warfarin embryopathy not as high as previously thought, especially if the mother achieves adequate anticoagulation on <5mg>
3. No data on low molecular weight heparin in this situation, so its use cannot be recommended.

The patient must be fully informed, and involved in deciding her mode of anticoagulation (medicolegal implications).

How do I manage the pregnant woman with dilated cardiomyopathy?

The management of a pregnant woman with dilated cardiomyopathy should be considered in terms of maternal risk, and risk to the fetus.

Maternal risk
This relates to the degree of ventricular dysfunction and the ability to adapt to altered haemodynamics. Risk and management
can therefore be discussed in relation to New York Heart Association (NYHA) functional class:

NYHA I-II
• Should manage pregnancy without difficulty (maternal mortality 0.4%)
• May require admission for rest and diuretic therapy
• Venous thrombosis prophylaxis with heparin for patients on bedrest

NYHA III
• At significant risk (maternal mortality for NYHA III-IV 6.8%)
• Planned hospital admission for rest, treatment of heart failure and monitoring
• Risk of deterioration in ventricular function which may not improve post-partum.
• Early delivery if heart failure progressive despite optimal inpatient management

NYHA IV
• Should be advised not to become pregnant. Therapeutic abortion should be considered.

Fetal risk
Fetal risk should be considered in terms of two factors:

1. Factors which put the mother at risk
2. Adverse effects from maternal drugs:
   
   • ACE inhibitors should be discontinued prior to conception because of the risk of embryopathy
   • Limited or unfavourable data on fetal effects of many antiarryhthmics
   • Beta blockers may be associated with maternal hypotension, and hence reduce placental perfusion. They may thus contribute to premature labour
   • Warfarin

Note that digoxin and verapamil are safe to use.

What do I do if an ICD keeps discharging?

Most patients who experience a single ICD shock do so for successful conversion of a malignant ventricular arrhythmia. However, it must be remembered that the default programming in an ICD is designed to maximise sensitivity at the expense of specificity. Consequently, a significant number of ICD shocks can be inappropriate.1 For example, multiple shocks in quick succession may indicate inappropriate therapy for an atrial arrhythmia or a problem with the rate sensing lead. For this reason, it is important to retrieve the stored data from the device using the appropriate programmer even after a single shock. Evaluation of events stored in the ICD memory shows intracardiac electrograms, far field electrograms and recorded intervals as well as the onset and stability of the tachycardia to determine appropriate or inappropriate therapy. Frequent episodes of ventricular arrhythmia will require antiarrhythmic drugs for suppression; sotalol is often effective as a first line drug in this situation.

The more common reason for multiple ICD shocks is recurrent ventricular arrhythmia. Patients experiencing “storms” of shocks should be adequately sedated, and monitored in a coronary care setting. Intravenous antiarrhythmic drugs should be used for rapid arrhythmia suppression. Electrolyte abnormalities should be sought and promptly corrected. Myocardial ischaemia has to be a serious consideration when recurrent ventricular fibrillation or polymorphic ventricular tachycardia is responsible for shocks. Most episodes of repetitive ventricular tachycardia respond to intravenous drugs such as lidocaine, procainamide or amiodarone allowing for oral loading with an antiarrhythmic agent in a more controlled fashion.

If it becomes apparent that shocks are being delivered inappropriately (e.g. atrial fibrillation with rapid ventricular rates or shocks with no apparent arrhythmia signifying a lead fracture) suppression of ICD function can be achieved by applying a magnet over the ICD generator. Unless specifically programmed to the contrary, one can temporarily disable the sensing circuit of most ICDs during the period that a magnet is held over the ICD generator and prevent unnecessary shock while awaiting availability of appropriate equipment for definitive ICD programming changes.

Other causes of inappropriate therapy include:

• Sinus tachycardia
• Lead fracture
• Diaphragmatic muscle sensing
• Electromagnetic interference

How do I follow up the patient with the implantable cardioverter defibrillator?

Follow up of the patient with an implantable cardioverter defibrillator (ICD) demands an integrated team approach. The cardiologist, technical staff and nurses involved should have a wide experience and knowledge of pacemakers and general cardiac electrophysiology. Current generation ICDs do not just shock the heart but provide complex regimens of tachycardia discrimination and anti-tachycardia pacing (ATP) as well as single and dual chamber bradycardia therapy.

Routine follow up may occur in a tertiary centre or a local hospital as long as the expert staff and necessary equipment such as programmers and cardiac arrest kit are available. Follow up should start before the device is implanted with an educational programme and support for the patient and immediate family members. Videos, information booklets and meeting other patients with ICDs may be of benefit.

No consensus exists as to the interval between routine follow ups. Previously the patient had to return every month or two to have a capacitor reform. This is not now necessary, as all modern ICDs will undertake this automatically. With most current devices a 3 to 6 month interval is usual but treat each patient according to their individual circumstances.

Good management of the ICD should aim to achieve the following objectives:

1. Monitor the performance of the therapy delivered by the device, look at the success and failure of the programmed regimes and any acceleration of arrhythmias. Use this information to optimise clinical effectiveness of the programming.
2. Measure necessary parameters of the ICD and leads to ensure correct function. These should include lead impedance, shock coil impedance (if possible non-invasively), battery voltage, charge time, R and P wave amplitudes as well as pacing thresholds.
3. Review the intracardiac electrograms to ensure no inadvertent sensing of noise or other interference.
4. Maximise device longevity by safe and effective reprogramming of parameters.
5. Minimise the risk of complications occurring both from inappropriate therapy delivered to the patient and those associated with wound and pocket infection.1
6. Anticipate the elective replacement of the device and plan for this eventuality.
7. Provide a support structure for the patient and their family including advice, counselling and education. Some centres provide a formal patient support group; there are both positive and negative views on this practice.

How do I manage the patient with an ICD?

An implantable cardioverter defibrillator (ICD) serves as prophylaxis against sudden collapse and death from rapid ventricular arrhythmias. In general, all ICDs sense the heart rate and provide anti-tachycardia pacing or deliver synchronised (cardioversion) or unsynchronised (defibrillation) shocks. Some of the modern ICDs also incorporate dedicated pacing function; patients with heart block or sinus node disease may be dependent on these devices just like any patient with an implanted cardiac pacemaker. Like pacemakers, ICDs have to be checked by telemetric interrogation at periodic intervals to confirm integrity of the lead systems and proper function of ICD components including adequacy of battery voltage. Reprogramming of the various parameters that govern pacing, arrhythmia detection and therapy may be necessary from time to time. Such routine follow up, usually undertaken at established arrhythmia centres, should occur at 3 to 6 monthly intervals in the absence of major intercurrent events. Some issues specific to this group of patients can be summarised as follows:

1. Avoid rapid heart rates
In its basic form, arrhythmia detection algorithms of ICDs rely on a programmed heart rate threshold. Once this is exceeded for a defined period of time, the device may deliver therapy irrespective of whether the arrhythmia is of ventricular or supraventricular origin. In a ventricular-based ICD, the shock energy vector is designed primarily to encompass the ventricles. Consequently, atrial arrhythmias may fail to convert such that multiple inappropriate ICD shocks may result. Further, if antitachycardia
pacing is delivered in the ventricle for an atrial arrhythmia, ventricular arrhythmias may be provoked creating a pro-arrhythmic situation. The newer ICDs incorporate atrial sensing to improve arrhythmia discrimination but it must be remembered that any algorithm that improves specificity for ventricular arrhythmia will entail some loss of sensitivity. Cognisant of the above, it is imperative that atrial arrhythmias are adequately treated in these patients, particularly the paroxysmal form of atrial fibrillation that is commonly associated with rapid rates at its onset. Occasionally, RF ablation of the AV node is necessary. Beta adrenergic blockers should be an integral part of therapy in most ICD patients.

2. Recognise ICD—drug interactions
Antiarrhythmic drugs have the potential for interacting with an ICD in several ways. Drugs such as flecainide and amiodarone can increase pacing and defibrillation thresholds. In patients with a low margin of safety for these parameters, use of these drugs may result in failure of pacing or defibrillation. Secondly, these drugs can slow the rate of ventricular tachycardia below the programmed rate threshold for detection by the ICD; failure of arrhythmia detection can result. Some rarer interactions include
alteration of the T wave voltage by drugs or hyperkalaemia resulting in double counting and inappropriate shocks.

3. ICD wound management
As an implanted device, the system is susceptible to infections. Pain and inflammation of the skin over the ICD may herald an infective process. Similarly, unexplained fever, particularly staphylococcal septicaemia may indicate endocarditis involving the leads and/or tricuspid valve.

What are the indications for implantable cardioverter defibrillator (ICD) implantation and what are the survival benefits?

The class two indications for ICD implantation are:

1. Cardiac arrest presumed to be due to VF when electrophysiological testing is precluded by other medical conditions.
2. Severe symptoms attributable to sustained ventricular arrhythmias while awaiting cardiac transplantation.
3. Familial or inherited conditions with a high risk for lifethreatening ventricular tachyarrhythmia such as long QT syndrome or hypertrophic cardiomyopathy.
4. Non-sustained VT with coronary artery disease, prior MI, and LV dysfunction, and inducible sustained VT or VF at electrophysiological study.
5. Recurrent syncope of undetermined aetiology in the presence of ventricular dysfunction and inducible ventricular arrhythmias at electrophysiological study when other causes of syncope have been excluded.

What are the indications for implantable cardioverter defibrillator (ICD) implantation and what are the survival benefits?

Studies in the early 1980s showed that recurrence rates were high for patients presenting with a malignant arrhythmia unrelated to myocardial ischaemia or infarction. Survivors of cardiac arrest, those presenting with sustained monomorphic VT and unexplained syncope in the presence of heart disease clearly are patients at high risk for sudden cardiac death. A series of clinical trials completed in the recent past have confirmed the uniform survival benefit from ICD therapy in such patients (AVID, CASH, CIDS) when compared to therapy with amiodarone or sotalol. In the largest prospective randomised trial (Antiarrhythmics versus Implantable Defibrillators Trial – AVID trial), the ICD reduced mortality by 39% at 1 year and 31% at 3 years. Most patients randomised to the antiarrhythmic arm of the trial were treated with amiodarone.

With remarkable improvements in ICD technology allowing easier implantation, the ICD is being embraced increasingly and earlier in the course of cardiac disease. Attention has now turned to primary prevention of sudden death. For patients with asymptomatic non-sustained VT, there appears to be a clear survival benefit from ICD in the presence of a remote myocardial infarction, LVEF <40%, and inducible VT at electrophysiological study (MADIT, MUSTT). Interestingly, this benefit cannot be extrapolated to patients without nonsustained VT or inducible VT. The CABG patch trial that randomised patients with LVEF <36% and positive signal averaged ECG to ICD or not during elective bypass surgery failed to show a survival benefit. The role of the ICD in primary prevention of sudden death in non-ischaemic dilated cardiomyopathy is also unclear at this time. Clinical trials are in progress.

The benefit from an ICD appears to be greatest for patients with severe LV function and additive to conventional therapy with ACE inhibitors and beta adrenergic blockers. In the AVID trial for example, survival benefit with ICD was observed only when LVEF was less than 35%. Similarly, in the primary prevention trials, the mean LVEF was 30%. One could advance the argument that the ICD should be reserved for those with the worst LV function. Unfortunately, such patients have competing causes for mortality such as pump failure and electromechanical dissociation that are responsible for 50% of deaths. On the other hand, patients with little or no impairment of LV function and a single tachyarrhythmic event usually have late and rare recurrence leading to sudden death. An ICD can potentially restore them to near normal life expectancy in the absence of ongoing myopathic process. The long term studies requiring more than one life span of an ICD are not available to define the true value of
ICD therapy in such patients.

Although the ability of the implantable cardioverter defibrillator (ICD) to terminate potentially lethal ventricular arrhythmias is well acknowledged there is less consensus as to whom should receive an ICD. A good place to start is the American College of Cardiology/American Heart Association Practice Guidelines for Arrhythmia Devices.1 There are three classes of indications: class one, where there is evidence and/or general agreement that the treatment is beneficial, useful and effective; class two, where there is conflicting evidence or a divergence of opinion; and class three, where there is evidence and general agreement that a treatment is not useful or effective.

The class one indications for ICD implantation are:

1. Cardiac arrest due to VF or VT not due to a transient or reversible cause.
2. Spontaneous sustained VT.
3. Syncope of undetermined origin with clinically relevant, haemodynamically significant sustained VT or VF induced at electrophysiological study when drug therapy is ineffective, not tolerated or not preferred.
4. Non-sustained VT with coronary disease, prior MI, LV dysfunction, and inducible VF or sustained VT at electrophysiological study that is not suppressible by a class I antiarrhythmic drug.

The class two indications for ICD implantation are:

1. Cardiac arrest presumed to be due to VF when electrophysiological testing is precluded by other medical conditions.
2. Severe symptoms attributable to sustained ventricular arrhythmias while awaiting cardiac transplantation.
3. Familial or inherited conditions with a high risk for lifethreatening ventricular tachyarrhythmia such as long QT syndrome or hypertrophic cardiomyopathy.
4. Non-sustained VT with coronary artery disease, prior MI, and LV dysfunction, and inducible sustained VT or VF at electrophysiological study.
5. Recurrent syncope of undetermined aetiology in the presence of ventricular dysfunction and inducible ventricular arrhythmias at electrophysiological study when other causes of syncope have been excluded.

Who should have a VT stimulation study? What are the risks and benefits?

Contrary to conventional wisdom, a significant number of sudden arrhythmic deaths result from re-entrant ventricular tachycardia that occurs in patients with chronic heart disease in the absence of acute infarction. These arrhythmias can be safely studied in a controlled setting using electrophysiological testing. Programmed electrical stimulation of the ventricle (also termed VT stimulation studies) has a remarkable sensitivity for reproducing monomorphic ventricular tachycardia associated with infarct related myocardial scars and offers a fairly reliable means of identifying patients at risk for sudden death. Patients with LV dysfunction (LV ejection fraction <40%) who are inducible for monomorphic VT have a risk of sudden cardiac death of approximately 30% over the ensuing year.

The patients at highest risk for sudden death include those who have survived a cardiac arrest not occurring in the context of an acute infarction, and those presenting with sustained VT. These patients are best treated with implantable cardiac defibrillators. The role of VT stimulation studies in such patients is primarily to confirm the diagnosis and exclude focal ventricular arrhythmias or unusual supraventricular arrhythmias indistinguishable from VT that are amenable to RF ablation. Occasionally, suppression of VT inducibility with drugs such as amiodarone and sotalol may be an acceptable alternative to implantable cardioverter defibrillator (ICD) implant.

VT stimulation studies are more valuable for patients with severe heart disease and unexplained syncope. Such patients may have had a self-limiting arrhythmia causing their syncope. Inducibility of monomorphic VT is a fairly specific finding in this patient population especially if their heart disease is based on coronary artery disease. In addition, electrophysiological studies can unmask severe His-Purkinje conduction disease requiring pacemaker implantation. One major drawback of VT stimulation studies is the low sensitivity for ventricular arrhythmia in nonischaemic dilated cardiomyopathy. In these patients, if the clinical suspicion is high, a negative study may well represent a false negative. A second problem with VT studies is the uncertain reliability of induced polymorphic VT or ventricular fibrillation as end points. Recent data from subgroup analysis of the Multicenter Unsustained Tachycardia Trial (MUSTT) suggests that such arrhythmias may be just as important as monomorphic VT for predicting mortality in the face of severe LV dysfunction.

Perhaps the most important role of VT study is in primary prevention of sudden death. Two recent randomised trials have demonstrated conclusively that patients with depressed LV function and non-sustained VT (defined as three or more beats of VT at a rate >120bpm) will benefit from ICD implantation if they are inducible for sustained VT.1,2 Clinical trials are in progress to determine if ICD implantation would benefit patients with low LVEF and heart failure alone without resorting to an EP study. Pending their results, patients with LV dysfunction who manifest non-sustained VT should undergo VT stimulation studies to see if they would benefit from an ICD. This strategy appears to be cost effective.

The risks of invasive electrophysiological studies are related to venous (and rarely arterial) cannulation and from the arrhythmias induced. Injury to the vascular structures and venous thrombosis occurs rarely (less than 2%). Cardiac perforation from catheter placement is equally rare (0.4%); death from the procedure occurred in 0.12% in one study4 and underlines the importance of trained personnel and well equipped laboratories for these studies.

What percentage of patients will suffer the complications of amiodarone therapy, and how reversible are the eye, lung, and liver changes?

What percentage of patients will suffer the complications of amiodarone therapy, and how
reversible are the eye, lung, and liver changes? How do I assess thyroid function in someone on amiodarone therapy?

Amiodarone therapy is associated with a number of serious toxicities which primarily involve the lung, heart, liver or thyroid gland. The drug is also associated with a wide array of other side effects involving the skin, eye, gastrointestinal tract and neurologic system. Drug discontinuance rates with amiodarone are closely related to its daily dose. The table summarises the cumulative incidence of adverse reactions reported in two separate meta-analyses.

Eye, lung, and liver toxicity are all potentially reversible if amiodarone is discontinued early after the development of toxicity. However, cases of permanent blindness, death from liver failure and death from respiratory failure have been rarely reported with amiodarone.

There are no adequate predictors of pulmonary toxicity, and serial lung function studies are usually not helpful. Dose and duration of treatment are no guide to risk. Clinical suspicion must remain high, especially in the elderly or those with co-existent pulmonary disease.

Amiodarone has been implicated as a cause of both hyperthyroidism and hypothyroidism. Hypothyroidism is a predictable response to the iodide load presented by amiodarone. Two types of hyperthyroidism have been reported to occur with amiodarone. Type I amiodarone-induced hyperthyroidism occurs in patients with underlying thyroid disease such as Graves disease. The iodide load in these patients accelerates thyroid hormone synthesis. Type II amiodarone-induced hyperthyroidism
occurs in patients with normal thyroids. Hyperthyroidism results from a direct toxic effect of amiodarone causing a subacute destructive thyroiditis with release of preformed thyroid hormone. Patients receiving amiodarone should have thyroid function evaluated at periodic intervals. A low TSH is indicative of hyperthyroidism, but does not distinguish between Type 1 and Type 2 hyperthyroidism. Radioactive iodine uptake may be low normal or elevated in Type 1 hyperthyroidism but is very low or
absent in Type 2 hyperthyroidism. Interleukin-6 levels are normal or moderately increased in Type 1, but markedly increased in Type 2 amiodarone-induced hyperthyroidism. In addition, colour flow Doppler ultrasound shows an absence of vascularity in Type 2 amiodarone-induced hyperthyroidism.

Amiodarone-induced hypothyroidism is characterised by an elevated TSH. Treatment of amiodarone-induced hypothyroidism is indicated if the free T4 is low or low normal and the TSH is greater than 20 microIU/ml.

As a complication of therapy, hyperthyroidism is more common where dietary iodine intake is low, whilst the reverse is true in areas of high intake.4 In patients with hyperthyroidism in whom amiodarone therapy is still warranted, thought should be given to concomitant treatment with carbimazole.

How do I investigate the relatives of a patient with sudden cardiac death?

In patients aged over 30 years by far the commonest cause of sudden cardiac death is coronary disease (80%). In patients younger than this, inherited disorders play a major role, with hypertrophic cardiomyopathy accounting for approximately 50% of these deaths. Although perhaps not entirely representative of the general population, the most systematically collected data on sudden death in young people comes from athletes. Aortic root dissection and arrhythmias due to accessory pathways and long QT syndromes may also be causative. A specific diagnosis in the deceased should be pursued by means of expert examination of the postmortem heart if available and attempts to obtain antemortem electrocardiograms and other investigations.

In general first-degree relatives should undergo history, physical examination, 12-lead electrocardiography and 2-D echocardiography. Other investigations may also be performed depending on the suspected cause of death, such as exercise testing in suspected long QT syndrome. In the case of a suspected inherited condition, if both parents of the deceased can be evaluated and found to be free of abnormalities, the condition causing death is likely to have been sporadic and the chances of siblings being affected are low. However, this inference must be tempered by the realisation that some inherited conditions (including hypertrophic cardiomyopathy) may be associated with incomplete penetrance. Extended pedigree analyses have demonstrated that occasionally apparently unaffected individuals, termed “obligate carriers”, carry the mutation. A follow up strategy after an initial negative evaluation is empirical, and depends on the age of the person, the level of anxiety and the nature of the suspected condition.

How do I assess the patient with long QT? Should I screen relatives, and how? How do I treat them?

Patients affected by the congenital long QT syndrome (LQT) are often first assessed when syncope, documented ventricular arrhythmia or aborted cardiac arrest affects them or a family member. The diagnostic cut-offs (<100%>0.46 sec (children <16>0.45 sec (adult males), and >0.47 sec (adult females), after drug induced QT prolongation has been excluded. T wave morphology should also be carefully examined, in particular for high takeoff, late onset, broad base, bifid morphology with humps, and beat-by-beat alternating polarity (T wave alternans). In several LQT variants, sinus bradycardia is an additional common feature. Holter monitoring should be performed to exclude repetitive ventricular arrhythmias of the torsade de pointes type. Family screening by 12-lead ECG of all first-degree relatives is mandatory in order to have a definite diagnosis of hereditary LQT. In Romano-Ward syndrome (1/20,000 births: autosomal dominant transmission with >90% penetrance), 50% of offspring of one affected parent are predicted to be similarly affected.

Six associated genetic loci (on chromosomes 3, 4, 7, 11, 21, 22) have been identified, of which four relate to genes that encode cardiac ion-channel proteins. Several mutations have been described for each gene. Although only 50% of all LQT affected families can be linked to one of these genes, genetic screening is 100% accurate amongst these, and can provide a definite diagnosis in phenotypically borderline cases.

Medical therapy should be promptly started in symptomatic LQT patients, and beta blockers are currently the first choice, with the occasional need for pacemaker implantation. However, recent evidence suggests that in symptomatic cases with aborted cardiac arrest, automatic implantable cardiac defibrillator (ICD) implantation, in addition to beta blocker therapy, is probably indicated. In patients who do not respond to the abovementioned measures, high cervicothoracic sympathectomy might be beneficial. Currently, there is no consensus regarding the need for therapy in asymptomatic patients, unless their phenotype is exceedingly abnormal. Gene-specific medical therapy is currently being investigated.

How do I treat torsade de pointes at a cardiac arrest?

Consideration of the electrophysiological disturbances predisposing to the development of torsade de pointes provides a logical approach to management. Experimental and clinical evidence implicates abnormal prolongation of cardiac action potential as a critical factor. Under these conditions early after-depolarisations may occur and lead to repetitive discharges (“triggered activity”).

Drugs that prolong cardiac action potential and are associated with torsade include antiarrhythmic agents of class Ia and III, tricyclic antidepressants, phenothiazines, macrolide antibiotics, certain antihistamines and cisapride. Hypokalaemia and hypomagnesaemia are well recognised causes of torsade although the evidence for hypocalcaemia is less convincing. Bradycardia – either sinus or due to atrioventricular block – is an important contributory factor.

In the setting of cardiac arrest torsade should be managed with synchronised DC cardioversion which is almost always successful in restoring sinus rhythm. However, additional measures will be necessary to prevent recurrence. These measures are aimed at shortening cardiac action potential duration. The heart rate should be increased. Atropine has the advantage of rapid availability and ease of administration. Where the bradycardia is due to atrioventricular block atropine is unlikely to increase the ventricular rate. Transvenous ventricular pacing should be established rapidly although it is almost certainly wise to stabilise the patient first with an isoprenaline infusion (at a rate of 1-10micrograms/min, titrated against the heart rate) or external cardiac pacing. There is experimental and clinical evidence to support the use of intravenous magnesium in the acute treatment of torsade. A dose of 8mmol (administered over 10-15 minutes) has been shown to abolish torsade in the majority of patients although a second dose may be necessary. There is no evidence to support the use of either intravenous potassium or calcium. The serum concentration of these electrolytes is frequently disturbed as a result of cardiac arrest per se and a reasonable strategy would be to obtain a formal laboratory measurement after a period of haemodynamic stability and to correct as necessary. Ventricular pacing should be maintained and the ECG monitored while the factors predisposing to the development of torsade are considered and corrected. There is no role for conventional antiarrhythmic drugs in the management of torsade de pointes: on the contrary many antiarrhythmics may aggravate the situation.

What do I do about non-sustained ventricular tachycardia on a 24 hour tape?

The term non-sustained ventricular tachycardia (VT) is used conventionally to describe salvos lasting a minimum of four consecutive ventricular beats and a maximum of 30 seconds in the absence of intervention. The concerns are that the non-sustained VT may itself cause symptoms of palpitation, presyncope or syncope and that the arrhythmia may persist or degenerate into ventricular fibrillation. The finding of non-sustained VT on a 24 hour tape should prompt the following questions: firstly, is there evidence of underlying heart disease; secondly, what is the morphology of the VT; thirdly, what are the patient’s symptoms?

An arrhythmia is usually although not invariably a sign of underlying heart disease. This is an important consideration because treatment of the underlying condition, where possible, is likely to be more effective than antiarrhythmic drug therapy both in terms of preventing the arrhythmia and improving prognosis. Conversely, if treatable underlying heart disease remains untreated then antiarrhythmic drug therapy is unlikely to be successful.

The morphology of the VT may help to guide management: for example if torsade de pointes is observed then management will focus on adjustment of drug regimes and treatment of electrolyte deficiencies and bradycardia. The finding of monomorphic VT might suggest the presence of a re-entrant circuit or automatic focus that may be amenable to mapping and modification or ablation. Non-torsade polymorphic VT is typically seen in the context of heart failure and is seldom reliably induced by electrophysiological study or amenable to radiofrequency ablation.

There is little evidence that antiarrhythmic drug therapy alters prognosis in patients with non-sustained VT. This may reflect a lack of efficacy and/or toxicity of currently available antiarrhythmic agents. Another explanation is that non-sustained VT is frequently a marker of underlying heart disease, which itself determines prognosis. There is evidence that implantable cardioverter-defibrillators (ICDs) may improve the prognosis of patients with poor left ventricular function, asymptomatic nonsustained VT and inducible, non-suppressible VT following myocardial infarction. However, many important questions remain about the prophylactic implantation of ICDs in such patients. The decision to implant is easier if there is a history of presyncope or syncope.

What do I do if a patient has a pacemaker and needs cardioversion?

Patients with pacemakers often require cardioversion, particularly with the increasing use of pacing techniques in the management of paroxysmal atrial fibrillation.

Some centres reprogramme or inactivate pacemakers prior to cardioversion. The decision regarding this should be made on an individual basis, depending on the type of pacemaker, reason for implant, and pacing-dependency.

Patients needing cardioversion should have the paddles applied in a manner such that the electrical field is remote from the pacemaker electrical field. In practise the standard apex— sternum approach is safe with a pacemaker in the left shoulder region, although anteroposterior paddle positioning can be utilised. The lowest energy possible should be administered, and the pacemaker should be checked formally after the procedure as occasionally the pacemaker may change mode as a consequence of cardioversion. Efforts should be made to ensure that, during synchronised shock, the defibrillator is recognising the ventricular, and not atrial, pacing spike.

Modern systems have increasingly effective protection from external interference.

Can a patient with a pacemaker touch an electric fence? …have an MRI scan? …go through airport metal detectors? …use a mobile phone?

Pacemakers have increasingly sophisticated circuitry to prevent damage or interference from external magnetic interference.

Electric fences
Nobody should touch an electric fence but should electric shock occur it would be wise to have the system checked by formal interrogation in case electrical mode reversion has occurred.

Magnetic Resonance Imaging (MRI)
MRI poses potential problems for the pacemaker patient. Significant artefact would be obtained in regions close to an implanted pacemaker but more importantly the powerful magnetic fields might interfere with the device. Initial blanket denial of MRI imaging to the pacemaker patient has been tempered by small studies showing device safety under carefully controlled conditions. Extreme caution should be advised and expert opinion sought prior to planned MRI investigation.

Airport metal detectors
Airport metal detectors have the potential to interfere with pacing systems. Patients should produce their pacemaker registration cards to bypass busy security queues.

Mobile phones
Mobile phones have been extensively investigated in terms of interaction with implanted devices. Analogue phones do not interact with implanted devices but more modern digital devices have the potential to interfere with pacing systems when utilised within a field of 10–15 cm. Pacemaker patients with mobile phones are therefore advised to carry mobile telephones on the opposite side of the body from the site of the device implant and should hold the device to the opposite ear.

Who should have VVI pacemakers and who should have dual chamber pacemakers? What are the risks of pacemaker insertion?

Many pacing enthusiasts argue that there are very few indications for VVI pacing, perhaps confining its role to the very elderly with established atrial fibrillation and documented pauses. Dual chamber pacing (or more accurately physiological pacing which may include single chamber atrial devices) is the preferred mode in most common indications for pacemaker implantation. The British Pacing group published its recommendations in 1991.1 These have led to widespread if gradual change in British pacing practice. Physiological pacemakers can be recommended in sinus node disease on the basis of many retrospective studies and one prospective study.2 Ongoing prospective studies will clarify the true role of physiological pacing in the elderly with AV conduction disease. The British guidelines are similar to those in the United States. A more comprehensive guide to pacemaker implantation is given by the ACC/AHA joint guidelines which supply the level of evidence for each ecommendation and a comprehensive reference list.

Pacemaker implantation is a remarkably safe procedure. Mortality is minimal and occurs due to unrecognised pneumothorax, pericardial tamponade or great vessel trauma. Complications at implant are those of subclavian puncture, particularly pneumothorax, although these can be avoided if the cephalic approach is used. There is some long term evidence that the cephalic approach may avoid chronic lead failure in polyurethane leads due to subclavian crush injury. Haematoma requiring re-operation should occur in less than 1%. Infection leading to explant similarly occurs in approximately 1%. Acute lead displacement should be less than 1% for ventricular leads and 1–2% for atrial leads.

Should the patient with trifascicular disease be routinely paced? If not, why not?

Normal activation of the ventricles below the bundle of His occurs by way of three “fascicles” – the right bundle branch and the anterosuperior and posteroinferior divisions of the left bundle branch. Conduction block in two of the three fascicles is bifascicular block. Additional prolongation of the PR interval results in “trifascicular block” implying abnormal conduction through or above the remaining fascicle. The concern is that conduction will fail in the remaining fascicle, i.e. complete heart block will develop with a slow and unreliable ventricular escape rhythm. Potential consequences include syncope and death.

There have been no randomised trials of pacing vs no pacing in patients with chronic bi- or trifascicular block. Clinicians must therefore be guided by knowledge of the natural history of the condition without pacing, and expert consensus guidelines. The largest prospective study of patients with bi- and trifascicular block followed 554 asymptomatic patients for a mean of 42 months. The five year mortality from an event that may conceivably have been a bradyarrhythmia was just 6%, a figure that must inevitably include some non-bradyarrhythmic deaths. The five year incidence of complete heart block was also low at 5%. A prolonged PR interval was associated with a higher incidence of potentially bradyarrhythmic deaths but not with the development of complete heart block. An important finding of this study was a five year all cause mortality of 35% reflecting the high incidence of underlying coronary heart disease and congestive cardiac failure.

The available evidence would suggest that asymptomatic patients with trifascicular block should not be paced routinely. A history of syncope should prompt thorough investigation for both brady- and tachyarrhythmic causes. If intermittent second or third degree block is documented permanent pacing is indicated. If tachyarrhythmias are implicated then therapy is likely to include antiarrhythmic drugs, which may exacerbate AV block and prophylactic permanent pacing would seem wise. Bi- and trifascicular block are associated with a high incidence of underlying coronary heart disease and heart failure. Attention should therefore be directed towards the detection of these conditions and the use of therapies known to improve their prognosis.

What are the chances of a 24 hour tape detecting the causes for collapse in a patient? What other alternative monitoring devices are now available?

Syncope is a common medical problem accounting for up to 6% of emergency medical admissions. In older patients presenting to casualty this may be as high as 20% when evaluated with a full cardiovascular work up. The annual recurrence rate is as high as 30%.1 Syncope due to cardiac causes is associated with a high mortality (>50% at 5 years) compared with 30% at 5 years in patients with syncope due to non-cardiac syncope and 24% in those with unexplained syncope.2 However, in the elderly, even “benign” syncope can result in significant morbidity and mortality due to trauma, anxiety or depression, which may lead to major changes in lifestyle or financial difficulties.

Syncope is often unpredictable in onset, intermittent and has a high rate of spontaneous remission making it a difficult diagnostic challenge. Thus even after a thorough work up, the cause of syncope may remain unexplained in up to 40% of cases.4 Prolonged ambulatory monitoring is often used as a first line investigation. Documentation of significant arrhythmias or syncope during monitoring is rare. At best, symptoms correlating with arrhythmias occur in 4% of patients, asymptomatic arrhythmias occur in up to 13%, and symptoms without arrhythmias occur in up to a further 17%.5–7 Prolonged monitoring may result in a slight increase in diagnostic yield from 15% with 24 hours of monitoring to 29% at 72 hours.

Patient activated external loop recorders have a higher diagnostic yield but do not yield a symptom-rhythm correlation in over 66% of patients, either because of device malfunction, patient noncompliance or an inability to activate the recorder.9,10 In addition such devices are only useful in patients with relatively frequent symptoms. In a follow up by Kapoor et al,11 only 5% of patients reported recurrent symptoms at 1 month, 11% at 3 months and 16% at 6 months. Thus this type of monitoring is likely to be useful only in a small subgroup of patients with frequent recurrence in whom initial evaluation is negative and arrhythmias are not diagnosed by other means, such as 24 hour ECG or electrophysiology studies.

The diagnostic yield from cardiac electrophysiology ranges from 14–70%. This variability is primarily dependent on the characteristics of patients studied, in particular the absence or presence of co-morbid cardiovascular disease.12 Thus despite the use of investigations such as head up tilt testing, ambulatory cardiac monitoring, external loop recorders and electrophysiological testing, the underlying cause of syncope remains unexplained and continues to pose a diagnostic problem.

The implantable loop recorder (ILR) is a new diagnostic tool to add to the strategies for investigation of unexplained syncope.12 It permits long term cardiac monitoring to capture the ECG during a spontaneous episode in patients without recurrence in a reasonable time frame. It should be considered in those who have already completed the above outlined investigations that have proved negative, and in those in whom the external loop recorder has not yielded a diagnosis in one month. The ILR is implanted under local anaesthetic via a small incision usually in the left pectoral region. It has the ability to “freeze” the current and preceding rhythm for up to 40 minutes after a spontaneous event and thus allows the determination of the cause of syncope in most patients in whom symptoms are due to an arrhythmia. The activation device, used by the patient, family member or friend freezes and stores the loop during and after a spontaneous syncopal episode. This is retrievable at a later stage using a standard pacemaker programmer. The ILR specifically monitors heart rate changes. Hypotensive syndromes including vasovagal syncope, orthostatic hypotension, post-prandial hypotension and vasodepressor carotid sinus hypersensivity may also cause syncope. An ability to record blood pressure variation in addition to heart rate changes during symptoms would be a very helpful and exciting addition to the investigation of people with syncope.

How should I investigate the patient with collapse? Who should have a tilt test, and what do I do if it is positive?

Investigation of a patient with collapse The history from the older patient may be less reliable, however a careful history often allows syncopal episodes to be classified into broad diagnostic categories. Elderly patients may have amnesia for their collapse. A witness history, available in only 40–60% of cases, can thus be invaluable. Witnessed features of prodrome (i.e. pallor, sweating, loss of consciousness or fitting) and clinical characteristics after the event can all help in building a diagnostic picture. Physical examination should include an assessment of blood pressure in the supine and erect position, a cardiovascular examination to look for the presence or absence of structural heart disease (including aortic stenosis, mitral stenosis, outflow tract obstruction, atrial myxoma or impaired left ventricular function) and auscultation for carotid bruits. The 12- lead electrocardiogram (ECG) remains an important tool in the diagnosis of arrhythmic syncope. Up to 11% of syncopal patients have a diagnosis assigned from their ECG. More importantly those with a normal 12-lead ECG (no QRS or rhythm disturbance) have a low likelihood of arrhythmia as a cause of their syncope and are at low risk of sudden death. Thus the history and physical examination can guide you as to the more appropriate diagnostic tests for the individual patient and would include the following:

• ECG
• 24 hour ECG
• 24 hour BP
• Carotid sinus massage – supine and erect (only if negative supine)
• External loop recorder
• Electrophysiological studies
• Head up tilt test
• CT head and EEG if appropriate
• Implantable loop recorder

Who should have a tilt test?
Kenny et al in 1986 were the first to demonstrate the value of head up tilt testing in the diagnosis of unexplained syncope.1 There is a broad group of hypotensive syndromes and conditions where head up tilt testing should be considered – patients with recurrent syncope or presyncope and high risk patients with a history of a single syncopal episode (serious injury during episode, driving) where no other cause for symptoms is suggested by initial history, examination or cardiovascular and neurological investigations. Tilt table testing may also be of use in the assessment of elderly patients with recurrent unexplained falls and in the differential diagnosis of convulsive syncope, orthostatic hypotension, postural tachycardia syndrome, psychogenic and hyperventilation syncope and carotid sinus hypersensitivity. What do you do if you make a diagnosis of vasovagal syncope on history and head up tilt test?

Which patient with a patent foramen ovale should be referred for closure?

A patent foramen ovale (PFO) occurs in approximately one quarter of the population. It is a vestige of the fetal circulation, with an orifice varying in size from 1 to 19mm, allowing right-toleft or bidirectional shunting at atrial level and the potential for paradoxical embolism. The development of better imaging techniques (e.g. transoesophageal echocardiography, contrast agents) and the fact that 35% of ischaemic strokes remain unexplained has generated interest in the potential for paradoxical thromboembolism through a PFO.

Studies of patients with cryptogenic stroke give a higher incidence of PFO (up to 56%)1 than in a control population, suggesting, but not proving, causality. Stroke due to paradoxical embolism involves the passage of material across a PFO, at a time when right atrial pressure exceeds left atrial pressure, to the brain. In one study the incidence of venous thrombosis as the sole risk factor for presumed embolic stroke in patients with PFOs was 9.5% and was clinically silent in 80% of patients,2 adding support to the concept of paradoxical embolism. The detection of venous thrombosis is not without difficulty and venous thrombi may resolve with time, such that a negative study does not exclude prior thrombosis. There is evidence that PFOs allow right-to-left shunting under normal physiological conditions, during coughing, straining and similar manoeuvres and especially in patients with raised right heart pressures and tricuspid regurgitation.

There are no completed prospective trials comparing aspirin, warfarin and percutaneous closure to guide management of patients with an ischaemic stroke presumed to be cardioembolic in origin. Opinion is divided in the case of a single ischaemic lesion on MR imaging and an isolated PFO – there is no evidence in favour of any particular strategy. Aspirin therapy is an uncomplicated option, and easier and safer than life-long warfarin. If there is evidence of more than one ischaemic lesion, no indication for warfarin (e.g. a procoagulant state), preferably a history of a Valsalva manoeuvre or equivalent immediately preceding the stroke and no alternative cause for the stroke then I would consider percutaneous closure, which has rapidly developed as a highly effective and technically straightforward procedure for closure of PFOs and many atrial septal defects.

What are the roles of transthoracic and transoesophageal echocardiography in patients with a TIA or stroke?

Approximately 80% of strokes are ischaemic in origin, of which 20–40% have a cardiac basis. TIAs have a cardiac cause in ~15% of cases. Common cardiac abnormalities associated with neurological events include atrial fibrillation, mitral valve disease, left atrial enlargement, left ventricular dilatation, prosthetic valve abnormalities and endocarditis. Clinical examination and simple tests (CXR and ECG) should indicate cardiac abnormality in these situations. The aim of echocardiography is to confirm the presence of important predisposing cardiac abnormalities and in younger patients, typically <50 years, to look for rare cardiac causes that might not be detected by other means. This latter group includes atrial septal aneurysm and patent foramen ovale (PFO) which, although somewhat controversial, are associated with an increased risk of stroke in patients without other detectable abnormalities.

Consequently, echocardiography is particularly useful in patients at both ends of the age scale. Older patients are more likely to have cardiac abnormalities that could give rise to stroke/TIA and young patients frequently have apparently normal hearts, but echocardiography (especially transoesophageal) may indicate the presence of an atrial septal aneurysm or PFO. The pick-up rate of transthoracic echocardiography is extremely low in patients with a normal clinical examination, CXR and ECG, making it a poor screening test. Conversely, the yield in patients with clinical abnormalities or an abnormal ECG/CXR is high and may give useful information for risk stratification beyond simply confirming a clinical diagnosis, for example left atrial size and the presence of spontaneous contrast.

Transoesophageal echocardiography should be reserved for “younger” patients (empirically <50 years) with unexplained stroke/TIA, for patients in whom the transthoracic study is unclear, and for older patients with repeated unexplained stroke/TIA. Transoesophageal echocardiography is particularly useful for looking at the left atrium, atrial septum, left atrial appendage, mitral valve and thoracic aorta, abnormalities of which may give rise to stroke/TIA. There is a tendency to over-report more subtle abnormalities (e.g. slight mitral valve prolapse) that may not be clinically relevant.

How sensitive are transthoracic and transoesophageal echocardiography for the detection of thrombus in the left atrium?

The ability of echocardiography to detect left atrial clot is determined by the sophistication of the equipment, the ease with which the left atrium and left atrial appendage can be scanned and the skill and experience of the operator. Historically, at best, the sensitivity of two dimensional transthoracic echocardiography for detecting left atrial thrombus has been of the order of 40–65%, with the left atrial appendage visualised in under 20% of patients even in experienced hands. This compared with a reported sensitivity of 75–95% for visualising left ventricular thrombi from the transthoracic approach. More recent data, from a tertiary referral centre using the new generation transthoracic echocardiography, suggest the left atrial appendage can be adequately imaged in 75% of patients and that within this group 91% of thrombi identified by transoesophageal echocardiography will also be visualised from the transthoracic approach. Although encouraging, the extent to which these figures can be reproduced using similar equipment by the generality of units remains to be established.

Available data for the sensitivity of transoesophageal echocardiography in detecting left atrial and left atrial appendage thrombus consistently report a high positive predictive value. The largest series of 231 patients identified thrombus ranging from 3 to 80mm in 14 patients: compared with findings at surgery this produced a sensitivity of 100%. But these findings need to be interpreted with considerable caution and are unlikely to be applicable to all users of the technique. The study was carried out in a tertiary referral centre with a particular interest and long-standing investment in the technique and the nine observers involved in reporting the data all had extensive experience. Nonetheless, transoesophageal echocardiography is undoubtedly the investigation of choice for imaging the left atrium and left atrial appendage.

How do I assess the risk of CVA or TIA in a patient with chronic atrial fibrillation and in a patient with paroxysmal atrial fibrillation?

Age is an important determinant of the risk of thromboembolism, and hence of transient ischaemic attack (TIA) and of cerebrovascular accident (CVA) in patients with atrial fibrillation. If the patient is aged less than 60 years, and has no evidence of other cardiac disease (such as coronary artery disease, valve disease or heart failure) the risk of thromboembolism is low (of the order of 0.5% per year). This risk is lower than the risk of a serious bleed if the patient is anticoagulated with warfarin (1.3% per year or higher depending on the quality of anticoagulation control). If the patient is older than the 60 years, or has evidence of other cardiovascular disease, the risk of thromboembolism increases.

In the Stroke Prevention in Atrial Fibrillation Study clinical features indicating a higher risk of thromboembolism were: age over 60 years; history of congestive heart failure within the previous 3 months; hypertension (treated or untreated); and previous thromboembolism. The more of these features present in a patient the higher the risk of thromboembolism. A large left atrium (>2.5cm diameter/m2 body surface area) or global left ventricular systolic dysfunction on transthoracic echocardiography also identifies patients at a higher risk of thromboembolism. Such abnormalities may not be suspected clinically and wherever possible echocardiography should be performed in patients with AF in order to determine more precisely the risk of thromboembolism.

Paroxysmal (as opposed to chronic) atrial fibrillation covers a wide spectrum of disease severity with the duration and frequency of attacks varying markedly between and within patients. Although the clinical trials of anticoagulation in patients with atrial fibrillation were inconsistent in including patients with paroxysmal atrial fibrillation, there was no evidence that such patients had a lower risk of thromboembolism than those with chronic atrial fibrillation. It is likely that as the episodes become more frequent and of longer duration that the risk approaches those in patients with chronic atrial fibrillation.

Are patients with atrial flutter at risk of embolisation when cardioverted? Do they need anticoagulation to cover the procedure?

Although common clinical practice and guidelines do not advocate routine anticoagulation of patients with atrial flutter undergoing cardioversion, there are no data to support this practice. Rather, recent studies suggest the prevalence of intraatrial thrombus in unselected patients with atrial flutter is significant and of the order of 30–35% (compared with 3% prevalence in a control population in sinus rhythm). The atrial standstill (or stunning) that has been described post-cardioversion of atrial fibrillation and is thought to be a factor in the associated thromboembolic risk has also now been described immediately post-cardioversion of patients with atrial flutter. Although some authors argue that the stunning post-cardioversion of atrial flutter is “attenuated” compared with the response in atrial fibrillation, the thromboembolic rate associated with cardioversion of atrial flutter in the absence of anticoagulation argues against this. Indeed, the thromboembolic rate appears to be comparable with the early experience of cardioverting atrial fibrillation. Furthermore, atrial flutter is an intrinsically unstable rhythm which may degenerate into atrial fibrillation and certain patients alternate between atrial fibrillation and atrial flutter.

Like atrial fibrillation, atrial flutter may be the first manifestation of underlying heart disease and it is likely, though not yet proven, that the thromboembolic risks associated with both chronic atrial flutter and with cardioversion of atrial flutter vary with the extent of underlying cardiovascular pathology. Although existing data are limited, on current evidence we advise that patients with atrial flutter should be anticoagulated prior to, during and postcardioversion, in the same way as patients with atrial fibrillation.

What are the risks of elective DC cardioversion from atrial fibrillation?

There are relatively few recent published data on the risks of elective DC cardioversion. The risks include those relating to an, albeit brief, general anaesthetic which will reflect the overall health of the patient, and those relating to the application of synchronised direct current shock. The latter include the development of bradyarrhythmias (more likely in the presence of heavy beta blockade and especially where there is concomitant use of calcium channel antagonists) and tachyarrhythmias (more
likely in the presence of deranged biochemistry including low serum K+ or Mg++, and high levels of serum digoxin). These dysrhythmias may necessitate emergency pacing or further cardioversion and full resuscitation. Elective cardioversion of adequately assessed patients should only be undertaken by appropriately trained staff in an area where full resuscitation facilities are available. Following cardioversion high quality nursing care and ECG monitoring will be required until the patient has recovered from the anaesthetic and is clinically stable. Failure to observe these guidelines will likely result in higher
complication rates which on occasion includes death.

The other major complication of DC cardioversion is thromboembolism which can be debilitating and is sometimes fatal. There have been no randomised trials of anticoagulation but there is convincing circumstantial evidence that anticoagulation reduces the risk of cardioversion-related thromboembolism from figures in the order of 7% to less than 1%: anticoagulation does not appear to abolish the risk and this should be made explicit when informed consent is obtained from a patient. Patients with recent onset AF are not devoid of the risks of cardioversion-related thromboembolism and also require anticoagulation.

What factors determine the chances of successful elective cardioversion from atrial fibrillation?

Elective cardioversion should only be undertaken when the precipitant (e.g. hypoxia, ischaemia, thyrotoxicosis, hypokalaemia and hypoglycaemia) has been treated and the patient is metabolically stable. With this proviso, the success of cardioversion depends not so much on the ability to restore sinus rhythm (success rates of 70–90% are usual), but rather on the capacity to sustain sinus rhythm.

Cardioversion of unselected patients will result in consistently high rates of relapse: at one year 40 to 80% of patients will have reverted to atrial fibrillation. Early cardioversion, particularly in those patients in whom a clear trigger of atrial fibrillation has been
effectively treated and in whom there is little or no evidence of concomitant cardiac disease, is associated with the best long term outcome. This may reflect the finding (well described in animal models) that sustained atrial fibrillation modifies atrial electrophysiology
so that, with time, there is a predisposition to continued and recurrent AF. If early cardioversion is not feasible, then the extent of underlying cardiac disease may be a more important determinant of long term outcome than the duration of AF.

The presence of severe structural cardiac disease is associated with a high relapse rate and sometimes an inability to achieve cardioversion, e.g. severe ventricular dysfunction, markedly enlarged atria and valvular disease.

Certain categories of patients justify specific mention:

• Obese patients may be especially resistant to cardioversion from the external route but not necessarily using electrodes positioned within the heart.
• A proportion of patients with paroxysmal atrial fibrillation will eventually develop chronic atrial fibrillation: for many this provides a paradoxical reprieve from their symptoms. If cardioverted their propensity to atrial fibrillation remains and they are likely to relapse.
• The prognosis of patients with structurally normal hearts who develop atrial fibrillation as a result of thyrotoxicosis is excellent: once the thyrotoxicosis has been treated a high proportion revert to sinus rhythm and the remainder are sensitive to cardioversion with a relatively low relapse.

How long should someone with atrial fibrillation be anticoagulated before DC cardioversion, and how long should this be continued afterwards?

For years, the rationale for a period of anticoagulation prior to cardioversion was that the anticoagulation would either stabilise or abolish any thrombus, the assumption being that thromboemboli associated with cardioversion occurred when effective atrial contraction was restored, dislodging pre-existing thrombus. Furthermore, it was assumed that recent onset atrial fibrillation was not associated with left atrial (LA) or left atrial appendage (LAA) thrombus and could therefore be safely cardioverted without anticoagulation. Although this has become standard clinical practice it is not evidence-based and not without hazard. With anticoagulation most thrombus appears to resolve rather than to organise. In patients with non-rheumatic atrial fibrillation most atrial thrombi will have resolved after four to six weeks of anticoagulation but persistent thrombus has been reported. Left atrial thrombus is present in a significant proportion of patients with recent onset atrial fibrillation and the associated thromboembolic rate is similar to that found in patients with chronic atrial fibrillation. Furthermore, cardioversion itself is associated with the development of spontaneous contrast and new thrombus and, in the absence of anticoagulation, even those patients who have had thrombus excluded using transoesophageal echocardiography have subsequently developed symptomatic thromboemboli.

For most patients a period of 4 to 6 weeks of anticoagulation and a transthoracic echocardiogram prior to cardioversion will be sufficient. Patients at high risk of thrombus (e.g. those with cardiomyopathy, mitral stenosis or previous thromboembolism) should undergo a transoesophageal study prior to cardioversion. In certain patients there may be cogent arguments for minimising the period of anticoagulation. In these patients transoesophageal echocardiography can be undertaken and provided no thrombus
is identified will abolish the need for prolonged anticoagulation prior to cardioversion. However, all patients with atrial fibrillation need to be fully anticoagulated at the time of cardioversion and for a period thereafter.

The duration of post-cardioversion anticoagulation should be dictated by the likely timing of the return of normal LA/LAA function and the likelihood of maintaining sinus rhythm. If atrial fibrillation has been present for several days only, normal atrial function will usually be re-established over a similar period and intravenous heparin for a few days post-cardioversion is probably adequate. Where the duration of AF is longer or unknown a period of anticoagulation with warfarin for 1–3 months is advised (reflecting a slower time course of recovery of atrial function).

What drugs should I use for chemically cardioverting atrial fibrillation and when is DC cardioversion preferable?

Drugs are more likely to be effective when used relatively early following the onset of atrial fibrillation. However, when a clear history of recent onset atrial fibrillation has been obtained it is important to establish and treat the likely precipitants. In many instances this will allow spontaneous reversion to sinus rhythm. Important precipitants include hypoxia, dehydration, hypokalaemia, hypertension, thyrotoxicosis and coronary ischaemia. Whilst these precipitants are being treated rate control will usually be required. Short acting oral calcium channel blockers (verapamil or diltiazem) and short acting beta blockers titrated against the patients response are most effective in this setting and likely to facilitate cardioversion. Intravenous verapamil should be avoided. If a patient with new atrial fibrillation is haemodynamically compromised urgent cardioversion is required with full heparinisation. Similarly patients with fast, recent onset atrial fibrillation with broad complexes are probably best treated with early elective DC cardioversion with full heparinisation.

With the above provisos there is a role for chemical cardioversion. Amiodarone (which has class III action and mild beta blocking activity) given through a large peripheral line or centrally can be highly effective, though a rate slowing agent may also be needed. Intravenous flecainide (class I) can also be highly effective. Like other class I agents (quinidine, disopyramide and procainamide), flecainide is best avoided in patients with known or possible coronary artery disease and in conditions known to predispose to torsade de pointes. Digoxin has no role in the cardioversion of atrial fibrillation.

The highest likelihood of successful cardioversion in patients with chronic atrial fibrillation is with DC cardioversion following appropriate investigation and anticoagulation. It should be noted that cardioversion is generally safe during digoxin therapy, so long as potassium and digoxin levels are in the normal range.

Which patients with SVT should be referred for an intracardiac electrophysiological study (EP study)?

Which patients with SVT should be referred for an intracardiac electrophysiological study (EP study)? What are the success rates and risks of radiofrequency (RF) ablation?

The management of supraventricular tachycardia (SVT) has changed dramatically with the development of curative radiofrequency ablation (RF ablation). For most patients, the technique offers a clear alternative to long term antiarrhythmic drug therapy with its potential toxic side effects. Except for atrial fibrillation and atypical atrial flutter, most SVTs are amenable to RF ablation albeit with some variation in success rates depending on the arrhythmia mechanism.

AV nodal re-entrant tachycardia and SVTs mediated via accessory pathways are the easiest to treat with RF ablation with success rates that exceed 90%.1 Recurrence is rare occurring in less than 10%. Focal atrial tachycardias and re-entrant atrial tachycardias resulting from prior atrial surgical scars have lower success rates of about 80%. Even for the rare but troublesome atrial tachycardia that cannot be ablated, RF ablation of the AV node with permanent cardiac pacing is effective in alleviating symptoms and can reverse any tachycardia mediated cardiomyopathy. Atrial flutter of the classical variety use a single re-entrant circuit in the right atrium and typically require an isthmus of tissue between the inferior vena cava and tricuspid valve for maintenance of the arrhythmia. RF ablation to create conduction block in this isthmus is effective in preventing recurrence of atrial flutter in 80% of patients with negligible risks. Unfortunately some patients develop atrial fibrillation because both arrhythmias share common cardiac disease processes that act as substrates for the arrhythmia mechanism. Nonetheless, fibrillation is easier to manage with drugs and combination of flutter ablation and antiarrhythmic drug therapy is often successful in maintaining sinus rhythm.

In the adult patient with the symptomatic Wolff Parkinson White syndrome, it is now generally believed that RF ablation should be the treatment of choice. Recurrent arrhythmias associated with ventricular pre-excitation are difficult to treat medically and often require the use of antiarrhythmic drugs with potent pro-arrhythmic effects or organ toxicity (e.g. flecainide, amiodarone). The risk of AV block is remote (less than 1%) unless the accessory pathway is located close to the AV node in which case the risk is higher. In infants and young children, on the other hand, it is often worth deferring RF ablation if possible because there is a chance that ventricular pre-excitation may resolve over a few years.

In contrast to the above, arrhythmias such as AV nodal reentrant tachycardia often respond to acute or interval therapy with one of the more benign AV nodal blocking agents e.g. digoxin, beta blockers or calcium blocker. RF ablation should therefore be reserved for recurrent or troublesome arrhythmia. Situations that justify earlier RF ablative therapy include haemodynamic instability during episodes, intolerance of drugs, desire to avoid long term drug therapy or occupational constraints such
as in airline pilots. It is also worth bearing in mind that once a patient requires more than two drugs for prophylaxis, it becomes more cost effective to proceed to RF ablation. The risk of AV block during RF ablation for AV nodal re-entrant tachycardia is between 1 and 2%,2 and is dependent on the experience of the operator. In the younger patients, even this low rate of complication can be important considering life time commitment to cardiac pacing in the event of heart block.

The risk of RF ablation is primarily that of AV block as noted above. Other risks are those related to cardiac catheterisation and include vascular damage, cardiac tamponade, myocardial infarction, cerebrovascular or pulmonary embolism and rarely damage to the valve in left sided pathways. In experienced centres, the risk of serious complications is less than 1%.

Which patients with paroxysmal or chronic atrial fibrillation should I treat with aspirin, warfarin or neither?

Patients in whom the risk of thromboembolism is considered to be greater than the risk of a serious bleed due to warfarin should be considered for formal anticoagulation. In published clinical trials of anticoagulation the risk of stroke was reduced from 4.3% per year to 1.3% per year with anticoagulation. This equates to 30 strokes prevented for 1000 patients treated with warfarin for 12 months. Whether such benefit can be seen in routine practice depends not only on a careful decision for each patient regarding the risk of bleeding and the risk of thromboembolism, but also on the quality of monitoring the intensity of anticoagulation. The usual practice is to anticoagulate to a target INR of 2.5 (range 2–3), unless there is a history of recurrent thromboemboli in which case higher intensity anticoagulation may be necessary. In the clinical trials the risk of serious bleeding was 0.9% per year in the control group and slightly higher (1.3%) in those on warfarin. Risk factors for bleeding on anticoagulants include serious comorbid disease (such as anaemia, renal, cerebrovascular or liver disease), previous gastrointestinal bleeding, erratic or excessive alcohol misuse, uncontrolled hypertension, immobility, and poor quality clinical and anticoagulant monitoring.

Aspirin therapy is often recommended for elderly patients with atrial fibrillation on the basis that there is a lower risk of bleeding compared with warfarin. The likely benefits of aspirin are also less than those of warfarin. Further, the bulk of AF-associated stroke occurs in those aged >75 years, and the benefits of anticoagulation are not outweighed by the risks in high-risk elderly patients in whom monitoring is carefully carried out.1 Where warfarin is genuinely considered unsuitable (or is unacceptable to a patient), and the patient is at significant risk of thromboembolism, there is evidence that aspirin at a dose of 325mg per day reduces the risk of thromboembolism, but no evidence that lower doses are effective. The combination of fixed-dose low intensity warfarin with aspirin confers no benefit over conventional warfarin therapy in terms of bleeding risks and is less effective in preventing thromboembolism.

What drugs should be used to maintain someone in sinus rhythm who has paroxysmal atrial fibrillation? Is there a role for digoxin?

The natural history of patients with paroxysmal atrial fibrillation is that over a period of time (and often many years) there is a gradual tendency to an increased frequency and duration of attacks. A proportion of patients will develop chronic atrial fibrillation. Not all patients require antiarrhythmic drugs and the potential side effects and inconvenience of regular medication must be balanced against the frequency of episodes and symptomatology which vary markedly between patients. Triggers include alcohol and caffeine, ischaemia, untreated hypertension (which if aggressively managed can at least in the short term obviate the need for antiarrhythmics), thyrotoxicosis, and in a small proportion of patients vagal or sympathetic stimulation where attacks are typically preceded by a drop in heart rate or exercise respectively.

The most effective drugs are also those with potentially dangerous side effects. The risks of class 1 agents (such as flecainide, disopyramide and propafenone) in patients with underlying coronary artery disease are well recognised and are best avoided. In younger patients (where it is presumed the associated risks are proportionately less) they can be highly effective. Sotalol may be useful in some patients but adequate dosing is required to achieve class 3 antiarrhythmic activity and not all patients will tolerate the associated degree of beta blockade. Amiodarone can be highly effective but its use is limited by the incidence of serious side effects. Beta blockers and calcium channel blockers have no role in preventing paroxysms of atrial fibrillation but can help certain patients in reducing the rate and so symptomatology.

Despite the long-standing conviction of many clinicians that digoxin is efficacious in the management of paroxysmal atrial fibrillation it has been clearly shown that digoxin neither reduces the frequency of attacks nor produces any useful reduction of heart rate during paroxysms of atrial fibrillation. Furthermore a number of placebo-controlled studies designed to explore the possibility that digoxin might chemically cardiovert patients with recent onset atrial fibrillation have shown no effect of digoxin as compared with placebo. Hence there appears to be no role for digoxin.

Can a cardiac transplant patient get angina? How is this investigated?

Post-transplant cardiac denervation theoretically abolishes the perception of cardiac chest pain. However, some patients may develop postoperative typical anginal chest pain precipitated by exercise or by increasing heart rate. This has been associated with ECG evidence of ischaemia and coronary angiography has confirmed transplant associated coronary artery disease. Such symptoms, however, are usually described by patients who are more than five years following transplantation. Chest pain associated with coronary artery disease is uncommon in patients who are less than five years post-cardiac transplantation. Interestingly, recent evidence shows an absence of bradycardic response to apnoea and hypoxia in cardiac transplant recipients with obstructive sleep apnoea. It may be that prospective overnight polysomnography studies will identify parasympathetic re-innervation in this group.

The majority of patients with transplant associated coronary artery disease do not get chest pain. Presenting features include progressive dyspnoea with exertion or the signs and symptoms of cardiac failure. Cardiac auscultation may reveal a third or fourth heart sound or features of heart failure. The ECG may show rhythm disturbances or a reduction in total voltage (the summation of the R and S wave in leads I, II, III, V1 and V6). Transthoracic 2D echocardiography may reveal evidence of poor biventricular function. Most units do not advocate routine annual coronary angiography for asymptomatic patients, since the angiographic findings do not usually alter clinical managment. Furthermore, conventional coronary angiography does not always confirm the diagnosis; intravascular ultrasound may be more sensitive. The condition is frequently diffuse and distal and not usually amenable to intervention, e.g. with angioplasty, stent insertion or bypass surgery. In those patients who have a localised lesion, the disease may progress despite successful intervention. The majority of centres do not usually offer cardiac re-transplantation on account of shortage of donor organs and poor results attendant on cardiac re-transplantation. Therefore patients who develop this condition are usually managed medically.

What drugs do post-transplant patients require, and what are their side effects? How should I follow up such patients?

Following successful cardiac, cardiopulmonary or pulmonary transplantation, patients require life-long immunosuppressive therapy. Routine immunosuppression consists of cyclosporin-A and azathioprine, occasionally supplemented by corticosteroids.

Episodes of acute allograft rejection are treated with intravenous methylprednisolone therapy or occasionally antithymocyte globulin or OKT3. Other drugs used include tacrolimus, mycophenolate mofetil and cyclophosphamide. Early evidence suggests that mycophenolate mofetil (an antimetabolite drug) may be a useful alternative to azathioprine as maintenance postoperative immunosuppression. OKT3 is a monoclonal antibody raised in mice, which is directed against the lymphocyte CD3 complex. Although it is sometimes used for induction following transplantation it is now more frequently employed in the management of severe episodes of acute cardiac rejection.

Common complications following transplantation include allograft rejection and infection. It is of paramount importance to immunosuppress the patient to minimise the risk of allograft rejection, without over-immunosuppressing and thereby increasing susceptibility to opportunistic infection. For this reason, cyclosporin-A blood levels are regularly monitored postoperatively. Side effects include renal failure, hypertension, hyperkalaemia, hirsutism, gum hypertrophy and increased susceptibility to opportunistic infection and to lymphoproliferative disorders. Tacrolimus acts in a similar way to cyclosporin-A although it may be a more potent immunosuppressive agent. Although its side effect profile is similar, diabetes mellitus can be a complication. Azathioprine is an antimetabolite whose major side effects include bone marrow suppression and hepatic cholestasis. Occasionally pancreatitis can occur. Some patients who are intolerant of azathioprine are prescribed mycophenolate mofetil (which is less likely to cause bone marrow suppression) or cyclophosphamide. At the present time the precise role of tacrolimus and mycophenolate in post-cardiac and pulmonary transplant immunosuppression is unclear and requires further study. The side effect profile of corticosteroid therapy is well documented.

In addition to regular monitoring of drug levels and haematological (full blood count) and biochemical (renal and hepatic function, blood glucose) indices, one should be aware of drug interactions which may reduce or increase the levels or effectiveness of immunosuppressive agents. For example drugs which promote hepatic enzyme induction (e.g. anticonvulsants, antituberculous therapy) will reduce cyclosporin-A levels. Certain antibiotics (e.g. erythromycin) and calcium channel blockers (e.g. diltiazem) will increase cyclosporin-A levels. Similar interactions apply to tacrolimus. Non-steroidal antiinflammatory agents can potentiate nephrotoxicity when given with cyclosporin-A or tacrolimus. The dose of azathioprine has to be reduced by 70% if patients are also prescribed allopurinol.