Re-examining examination: misconceptions in clinical medicine

Do not accept what you hear by report, do not accept tradition, do not accept a statement because it is found in our books, nor because it is in accord with your belief, nor because it is the saying of your teacher. Be lamps unto yourselves.

– Gautama Buddha

Medicine is learned by the bedside and not in the classroom. Let not your conceptions of disease come from words heard in the lecture room or read from the book. See, and then reason and compare and control. But see first.

– William Osler

Introduction

The United Kingdom has a proud tradition of clinical teaching at undergraduate level. Furthermore, entry into higher training in most specialties is preceded by obligatory written and clinical examinations. This is at odds with, for example, the United States of America, where there is less emphasis on clinical examination, with most American Board exams purely a written test.

Nonetheless as with any branch of knowledge, some of what is taught is handed down as an almost unbroken lineage from teacher to pupil without critical appraisal of its value. Bad habits are often picked up in medical school or early in clinical training, and are then never eradicated. Clinical examination is still a cornerstone of our diagnostic armamentarium and improvement of our skills should be a lifelong pursuit.

This brief essay will concentrate on three traditional teachings that deserve re-examination.

‘Hepato-jugular reflux’

First described by William Pasteur in 1885 in the context of tricuspid incompetence 1 and recognized in 1898 by Rondot to be a valuable sign in patients with normal tricuspid valves, by 1925 clinicians realized that pressure anywhere over the abdomen – not just the liver – could elicit the sign. 2 In fact pressure applied directly to the liver is likely to be painful in congestive cardiac failure or Budd-Chiari syndrome owing to liver capsular stretch.

The ideal way to perform this test is to firmly press over the patient's peri-umbilical area for 30 seconds, while observing the column of blood in the internal jugular vein. The test is positive if there is a sustained elevation in the height of the jugular venous pulsation (JVP) of at least 4 cm (although some clinicians recognize a positive test result by observing the neck veins at the moment that pressure is released, thereby regarding a fall of more than 4 cm as positive). No increase, or a transient increase with return to normal height during the 30 seconds, counts as a negative result. 3

A positive test result indicates the inability of the right heart to handle an increased venous return, and this is found in subclinical right ventricular failure, tricuspid incompetence, tricuspid stenosis, constrictive pericarditis, cardiac tamponade and volume overload.

The test probably works by displacing splanchnic venous blood towards the heart. It has been suggested that in congestive cardiac failure (CCF) systemic venous hypertension makes the venous system inelastic, tight, and non-compliant. 4 In any such hydraulic system, pressure exerted upon smaller vessels (e.g. splanchnic) will be transmitted to larger vessels (e.g. cervical veins). A challenging alternative view is that in a normal patient the inferior vena cava (IVC) is a flaccid tube, which is compressed by abdominal pressure, thereby reducing venous return to the heart. Hence only if the IVC is already distended will compressing the abdomen create a pressure wave and raise the JVP, so some would say that abdomino-jugular reflux is per se a sign of pathology. 5 In fact as the physiology is still not completely understood it is probably safer to call this the abdomino-jugular test, and altogether omit the word reflux.

As with all tests of physical signs there is inevitable inter-observer variability. Nonetheless this test – performed correctly – has a 66% sensitivity and up to 100% specificity for distinguishing tricuspid from mitral incompetence. It has again a high specificity for diagnosing heart failure. 6

To perform the test in as objective a manner as possible the following pitfalls should be avoided:

Not maintaining pressure for 30 seconds to see if the elevation in JVP is sustained or merely transient;

Pressing on the liver itself, as in CCF the liver capsule will be distended and tender, hence hepatic pressure will be uncomfortable for the patient. Furthermore it is the large volume of splanchnic blood that you are trying to force back into the circulation, not just the small amount of blood contained in the liver;

Pedantic, but needless to say, this test should not be referred to as the hepato-jugular test. And the word reflex should not be applied to this test.

It is also important to remember that Budd-Chiari syndrome (hepatic vein thrombosis) can give a false-negative result. To avoid another potential cause of an erroneous result it is best to examine the right internal jugular vein while performing the test. This is because a tortuous aorta can occasionally cause kinking of the left innominate vein, leading to distension of the left internal jugular vein. 7

There are many interesting alternative ancillary approaches to the JVP, which include:

The von Recklinghausen's manoeuvre: The patient lies supine with one hand on bed and one hand over patient's own thigh. If the veins in the lower hand are distended and the upper hand veins are normal then the JVP is probably within normal limits. However if veins in both hands are distended then the JVP may be raised. Unfortunately this test lends itself to many false positives from local obstruction or peripheral venous constriction, and as far as this author knows there have been no studies of its specificity or sensitivity for detecting an elevated JVP;

The Gärtner manoeuvre: Inspect veins of patient's hand or ante-cubital fossa and slowly raise hand until visible veins disappear. The height above the manubrio-sternal angle at which the veins re-appear is the height of the JVP. 8 Again, the problems that beset the von Recklinghausen's manoeuvre can also interfere with this test.

Postural blood pressure

Despite a lack of evidence regarding postural (sometimes called orthostatic) changes in blood pressure, it is still near impossible to make any sort of referral regarding patients with either gastro-intestinal bleeding or a syncopal episode without the question of a ‘postural drop’ being posed.

Classical textbook teaching suggests that a postural drop of ≥20 mmHg systolic blood pressure (SBP) or ≥10 mmHg diastolic blood pressure is suggestive of significant hypovolaemia, 9 although there is no clear consensus on whether the patient should be standing for 1 minute or 2 minutes after a period of lying supine.

Upon standing, up to half a litre of blood shifts to the venous system in the lower body. Blood pressure is maintained due to an increase in cardiac output and systemic vascular resistance, hence a normal individual will experience an increase in heart rate of approximately 10 bpm, a slight drop in SBP, and a rise in diastolic blood pressure. All these keep the mean arterial pressure (MAP) constant. 10

Orthostatic hypotension occurs either when these compensatory mechanisms fail – either due to autonomic insufficiency, or due to such excessive vascular loss that compensation is inadequate.

Unfortunately 10% of subjects (and up to 30% of those over the age of 65 years) will have a postural decrease in SBP of 20 mmHg while being perfectly well. 10

In true acute hypovolaemia, the most valuable observation is either a postural rise in pulse of at least 30 bpm or postural dizziness so severe that it is impossible to complete the test. These criteria give a sensitivity of 97% and a specificity of 96% for acute blood loss of greater than 630 mL. 11 In experiments of phlebotomy on healthy subjects – and hence it can be hard to extrapolate theresults to a broader patient base – the above two criteria gave a sensitivity and specificity of 98% for loss of 1000 mL of blood, but unfortunately the test did not work if the subject had lost only 500 mL of blood, or if the subject only sat up instead of standing. 12 Supine tachycardia and/or hypotension is a specific but markedly insensitive sign of volume depletion: only 31% of subjects with 1000 mL venesection had either symptom, but recall that these subjects were healthy volunteers.

If a subject is unable to stand (pace due to dizziness, which is already significant), then it is often suggested that measurement of lying and sitting blood pressure be performed. In fact sitting has been shown to significantly reduce the degree of leg pooling, and hence further reduces the sensitivity of the test. 12

Excluding the above, the finding of a postural drop of 20 mmHg SBP has no proven value, being found just as frequently in patients before blood loss as after it. 13

To make postural tests valuable in light of the available evidence base, the following should be avoided:

Not checking the pulse – the 30 bpm orthostatic pulse rise is one of the only two reliable signs of volume depletion (the other being inability to complete the test due to dizziness). It should also be noted that absence of the pulse to rise in the context of a postural drop is suggestive of neurogenic rather than volume-depletion problems;

Not taking a drug history – calcium-channel blockers and beta-blockade will inhibit the diagnostically useful tachycardic response. Understandably there have been few studies examining the cardiovascular implications of rate control agents in the setting of acute blood loss;

Performing a lying and sitting BP – as discussed above, this has no value;

Performing the test at all if you think the patient is shocked – Primum non nocere. Shock needs early treatment rather than more and more tests to prove its presence;

Letting somebody else do the test for you if you cannot be sure that they will wait at least a minute and measure the pulse as well as BP.

The ankle jerk reflex

Thetis, the mother of Achilles, had wanted to make her son immortal, so when he was a baby she had taken him down into the Underworld and dipped him in the river Styx, making his body impervious to injury except for the heel by which he had been held. 14 Lamarckian inheritance 15 perhaps explains why modern humans are still vulnerable to neurological foot problems, even those due to disc herniation at S1 level, as well as other causes of S1 nerve root damage.

The ankle jerk reflex is one of the most difficult to elicit in a tidy and objective manner. Positioning can be difficult, and the actual technique is not always easy in a recumbent patient. Therefore it is understandable for a physician to struggle with the ankle jerk, then remind himself comfortingly with the old saw that it is absent in up to 50% of people over the age of 65 years anyway.

The afferent limb for the reflex is the posterior tibial nerve, with the efferent limb coming from segments L5 to S2, but mostly from S1. Hence an absent ankle jerk is an extremely useful sign of sacral (S1) pathology, with a sensitivity of 85% for root compression syndromes at S1 caused by, for example, intervertebral disc herniation at the L5–S1 interface. 16

There are many ways to elicit the reflex, and descriptions can be found in any of the main textbooks on clinical examination. However in a recumbent patient it is probably ideal to cross the legs in a ‘figure 4’ position with the ankle to be tested lying across and beyond the thigh (or shin) of the other leg. Another option is to passively dorsiflex the patient's ankle with their leg straight, and for the examiner to then hit their own fingers (‘mediate percussion’). 17

It is essential to avoid excessive forced dorsiflexion of the ankle when testing the reflex, as this will obviously obliterate any feedback from the test. A 90-degree angle will suffice, or enough to have the slightest of tension on the achilles. And it is key to ensure the calf muscle is relaxed.

Furthermore an ankle jerk is not absent until it has been tested with reinforcement from the Jendrassik manoeuvre. This latter involves the patient grabbing their wrists tightly together and pulling hard upon the examiner's command; the facilitation of deep reflexes lasts for a few seconds only 18 and the mechanism through which it works is still not clear. 19

The oft-quoted figures regarding so-called ‘absent’ ankle jerks are derived from unreliable sources. 20 According to studies from the 1960s, ankle jerks were absent in 15% of all patients, 21 and of this proportion nearly 80% had diabetes mellitus. Even so, among diabetics only one in three had an absent ankle jerk. The remainder had a wide variety of other pathologies, among them S1 radiculopathy, tabes dorsalis, myxoedema, Eadie syndrome, myotonic dystrophy, pernicious anaemia and alcoholic polyneuropathy.

In the elderly it has been suggested that up to 80% of those over the age of 80 years have absent ankle jerk reflexes 20 however a carefully conducted prospective study – using both the Jendrassik manoeuvre and mediate percussion – found ankle reflexes to be absent in only 6% of all elderly patients, of whom the majority had a known explanation. Ultimately only three of the 200 elderly patients studied (mean age 82 years) had missing ankle jerks that could be attributed solely to age. 21

So the next time a patient under examination is found initially to be without ankle jerks, take a moment, try mediate percussion with Jendrassik reinforcement and see if you can perceive even the slightest twitch.

To improve our ability as clinicians to detect subtle ankle reflexes, I suggest the following:

Good patient positioning, without too much or too little Achilles stretch;

Performing the Jendrassik manoeuvre before declaring a reflex absent;

Trying mediate percussion or another alternative technique for reflex elicitation, such as having the patient kneel on a chair, or perhaps best of all testing the reflex with the patient lying prone and foot gently dorsiflexed;

Using the correct tendon hammer. Debatable, as no studies have ever shown superiority of one hammer over another. Nonetheless this author favours use of the classic Queen Square reflex hammer (a long pliable plastic rod with a circular disc at the end).

Conclusion

This article has briefly reviewed a few areas where conventional teaching and understanding of physical examination may be improved. There are, of course, a multitude of other areas where received wisdom is not as helpful as it could be; examples include assessment of proptosis, objective differentiation between ‘fine’ and ‘coarse’ crackles on auscultation of the chest, correct use of the bell of the stethoscope, and how it is possible to measure the blood pressure in a patient with atrial fibrillation. For now I leave these to the thought and consideration of the reader.

Footnotes

DECLARATIONS

Competing interests None declared

Funding None

Ethical approval Not applicable

Guarantor JMB

Contributorship JMB is the sole contributor