The following case history illustrates the importance of surreptitious cardiovascular examination in children, as this severe complication could have been avoided if her heart condition had been detected sooner and interventions performed at an earlier stage.

Presenting Complaint: Miss B, 14 years-old, was admitted with sudden onset of right-sided weakness and paraesthesia. Prior to this, she had been well.

History of Presenting Complaint: On the morning of her admission, she had been well, had dressed and gone to school as normal. While swimming at school, she noticed a progressive weakening of her right side. This worsened over the next 30 minutes and she began to feel paraesthesia in her right arm and leg. Neither prior to nor during these proceedings did she complain of headache, visual disturbance, loss or decrease in consciousness or any change in smell or hearing. She remained fully alert and orientated throughout.

Past Medical History: Apart from a history of 3–4 urinary tract infections (UTIs) in early childhood, her past medical history was unremarkable. She had no previous surgical history.

Family History: Her mother has suffered two deep vein thrombi and is on life-long warfarin therapy. Her maternal uncle died young of a stroke following several episodes showing symptoms of transient ischaemic attacks.

On Examination: She was alert, orientated and hydrated. She was hypertensive with a BP of 169/90. Power and sensation were decreased on the right and there was some evidence of facial asymmetry. A systolic murmur at the left sternal edge of grade 3–4/6 intensity was noted. This murmur radiated to the axillae and the carotids. The apex was noted to have a thrusting quality. Femoral pulses were impalpable.

Evolution: CT showed a haematoma close to the left basal ganglia. She was subsequently referred to the neurosurgical team at the Royal Manchester Children's Hospital (RMCH). An echocardiogram performed at RMCH revealed severe coarctation of the aorta. Further imaging by CT showed that the ascending aorta was dilated at 16 mm. The narrowing started just distal to the origin of the left subclavian artery where the maximum diameter of the thoracic aorta was noted to be 3–4 mm. Post stenotic dilatation of the aorta was also seen; here the diameter was found to be 18 mm. In addition, multiple collateral vessels (aorta–pulmonary vessels) were noted.

Coarctation of the aorta may not be picked up in the first few days of life as the four limb pressures and femoral pulses often remain normal until the ductus arteriosus closes. However, this case highlights the importance of how a full physical examination performed at any of her attendances for the UTIs may have led to the diagnosis being made sooner and an intervention performed before such a catastrophic complication occurred.¹ Current recommendations are that coarctation of the aorta is corrected surgically when patients are under 3 years of age.²

Heart murmurs are caused by turbulent blood flow but not all are secondary to structural heart lesions.³ The converse is also true that not all children with congenital heart disease (CHD) will have murmurs.¹ The prevalence of heart murmurs in children is quoted at 30–100%. This large variation reflects the wide variation of study sizes, study populations, experience of the examiner and the nature of the study (retrospec-tive/prospective).^3–5 These murmurs can be broadly divided into two categories – innocent and pathological. Pathological murmurs (see Table 1) arise from turbulent flow secondary to structural abnormalities. Innocent murmurs represent part of the normal spectrum of heart sounds; no structural abnormalities underlie these murmurs. They result from turbulence in structurally normal blood vessels, heart chambers or around the heart valves.^3,5 Although most heart murmurs are innocent, 5–8 in every 1000 children born will have a congenital heart lesion many of which are first diagnosed following detection of a heart murmur. About half of these lesions are insignificant and do not cause significant morbidity when managed properly. However, the other 50% can cause significant morbidity and may lead to death if not treated effectively and promptly.^1,6–8 The absence of abnormal signs including heart murmurs does not guarantee that the diagnosis of CHD can be excluded.^8,9 Children with CHD often need early medical or surgical interventions and can rapidly deteriorate if their lesions are not detected promptly.^2,5–6

Paediatricians are not expected to make a precise diagnosis as to the exact anatomical or physiological cause of each murmur they hear. They are required to have a high index of suspicion for CHD¹ and should be able to differentiate between innocent or pathological murmurs and appropriately refer those in need of further assessment. To do this, they must perform a full cardiovascular assessment as summarised in Table 2.¹¹ In particular, although it is not a diagnosis of exclusion, the following negative findings should be sought before diagnosing a murmur as innocent: no symptoms, no cyanosis, no signs of heart failure (increased respiratory rate/hepatosplenomegaly/growth failure), no dysmorphic syndrome associated with CHD, no other signs found on clinical examination. The diagnosis of innocent murmurs is somewhat simplified by the fact that the child is always asymptomatic but care must be taken to exclude the following congenital heart defects that may also be asymptomatic: small VSD and mild pulmonary and aortic stenosis.^1,5

   To summarise, patients needing referral to paediatric cardiology can be divided into three categories:

1. Those infants/children with:

• symptoms of heart failure, impalpable/decreased pulses and cyanosis.

• family history of CHD or sudden death.

2. Children who have murmurs of grade 3 intensity or greater or who have additional signs suggestive of a pathological murmur.

3. Those parents who are still anxious despite efforts to reassure them as to the innocent nature of the murmur.⁵

A number of studies concur that ECG and CXR are not usually necessary in the evaluation of heart murmurs that are thought to be innocent.^5,15–18 Tables 5 and 6 illustrate the more common of these characteristic findings. In addition, if these tests are solicited, it is important that they are reported by a clinician experienced in the wider range of normal values found in children and not by a computer program/adult technician.⁵ For example, being able to differentiate the normal right ventricular dominance seen in the first few months of life from right ventricular hypertrophy (RVH).¹⁴ It is important to remember the epitaph that 'the electrocardiogram may be normal in the patient with important heart disease and abnormal in patients with structurally normal hearts'.¹⁴

A high proportion of paediatric cardiology appointments are for the evaluation of heart murmurs referred by paediatricians. The principal aim of these consultations is to establish the cause of these murmurs and identify any pathological or structural cardiac lesions that may be present in a cost-effective manner. In addition, any necessary medical or surgical interventions can be arranged and antibiotic prophylaxis discussed. They also provide information to allay parental anxiety in children found to have murmurs whether they are innocent or pathological in nature.^11,19

A number of studies have shown that paediatric cardiologists have a high degree of sensitivity (92–96%) and specificity (94–95%) at differentiating between innocent and pathological murmurs by their clinical skills alone. These figures are even higher when echocardiography is employed (97% and 98%, respectively).^11,20,21

The British Paediatric Cardiology Association recommends that paediatric cardiologists see one patient every 20 minutes to allow full consultation and examination to be performed. The tertiary paediatric cardiology services in our centre are currently so overburdened that they are seeing 4 patients every 20 minutes and the majority of the cases in our age cohort are definitively diagnosed as having innocent flow murmurs.

Every child referred to the service receives an echocardiogram during their appointment. The current misperception by parents and paediatricians alike of cardiologists as gatekeepers of access to echocardiography is thought to underlie these high referral rates.^19,22 Echocardiography is beyond doubt the most reliable investigation for the exclusion of structural cardiac lesions. However, it is also abundantly clear that not all of the 50–100% of children who will at some stage have an innocent murmur need an Echo.⁵ Echo should, therefore, be reserved for those patients in whom the paediatrician has genuine suspicion of a cardiac lesion. Innocent murmurs can be evaluated in a cost-effective and highly reliable manner by clinical evaluation by a paediatric cardiologist alone. Obviously, there are exceptions to this rule, undue parental anxiety and positive family histories being the most commonly mentioned.^{5,11,12,16,20–25}

Many centres use sedation to improve the quality of the images captured by the scanner. With personnel shortage and time constraints in our centre, most echocardiographic imaging is performed without sedation. This means that it is very difficult to assess accurately some uncooperative children who must then be called back for follow-up.

In summary, it is clear that this service is overburdened. The following three proposals may help to reduce this burden:

1. Improve the confidence of paediatricians at detecting and differentiating between heart murmurs so that they refer less murmurs that they are convinced are innocent in nature.

2. Ensure that parents are informed by the referring doctor that their child may not have an echocardiogram when they have their appointment in paediatric cardiology as it may not be necessary.

3. Train and later evaluate the ability of paediatricians to perform echocardiography so that they can quickly assess murmurs locally and only refer those in whom there are concerns about the aetiology of the murmur.

To our knowledge, the ability of paediatricians to differentiate between innocent and pathological murmurs by clinical skills alone has not been assessed. The aim of this audit was, therefore, to assess the sensitivity, specificity and positive and negative predictive values of paediatricians at predicting the nature (pathological or innocent) of heart murmurs in children.

The audit was approved by the Department of Clinical Effectiveness at the Royal Bolton Hospital.

This audit was performed by the retrospective review of case notes of all new referrals to the Paediatric Cardiology Outpatient Clinic at the Royal Bolton Hospital for the evaluation of cardiac murmurs. Children were excluded if they had dysmorphic features associated with cardiac disease or if they had a previously diagnosed heart condition/lesion.

The notes of 41 patients, aged 1 month to 5 years, who attended this clinic between May 2004 and May 2005 as new referrals for the evaluation of heart murmurs were reviewed. Mean age, 35 months; median age, 28 months; and an age range of 3–56 months.

All children referred during the study period were seen by a single paediatric cardiologist with 9 years' experience. The echocardiography was performed using the Philips ultrasound M252404 (Bothell, WA, USA, 98041-3003). Images were obtained in all standard view and using pulse, colour and continuous Doppler images to identify any cardiac defects.

The outcome measure was comparison of the diagnostic impressions made by the general paediatrician compared with the definitive diagnosis made by the paediatric cardiologist and echocardiography – considered to be the gold standard.

Two-way table analysis of the initial and definitive diagnosis was employed to assess the sensitivity, specificity, positive and negative predictive values of the paediatricians at differentiating between pathological and innocent murmurs.