|The Art and Science of Percussion|
When you perform a physical examination of a fellow physician, you know that there are two people in the room who can do what you are doing. It makes you think.
Recently, I was examining one of our house officers. When I came to the pulmonary examination, I routinely began to percuss the lung fields. As I went through my usual maneuvers, the house officer asked, "Do you ever learn anything when you do that?" I told a couple of anecdotes about great findings I had made using percussion. The house officer seemed to be satisfied, but the question was unsettling. Why do I carry out this archaic procedure? Is it a waste of time? On days when my waiting room is filled with patients and time is of the essence, such small things can make a difference.
I thought about the radiologist who installed a stethoscope in a museum display case to make the point that auscultation of the lungs was out of date in an era of chest radiography. What, then, of this era of computed tomography and magnetic resonance imaging? And if auscultation is outdated, what about percussion?
There are certainly ways to identify intrathoracic pathology that are far more accurate than percussion. Nevertheless, percussion is sometimes indispensable. When a thoracentesis is planned, the clinician always confirms by percussion the upper margin of the fluid layer to be tapped, even if decubitus films and the results of ultrasound study are already available. Before abdominal paracentesis, the margins of an ascetic collection are likewise percussed. In the emergency room, the lungs are percussed to identify tympany associated with acute pneumothorax. During endotracheal intubation, the lungs are percussed to be sure that the tube is not in the right main stem bronchus. In each instance, clinicians rely on percussion to verify the results of chest x-rays and the still newer imaging technologies. We want to find out for ourselves; we are more trustful of test results when they confirm what we have discovered on our own examination. Percussion is also indispensable when a chest film is impractical or impossible to obtain, when the CT scanner or MRI machine is "down," or when an answer is needed immediately.
Percussion and auscultation were promulgated in Paris at about the same time: While Rene Laennec was experimenting with a long tube that permitted him to hear the heart sounds without actually placing his ear on the chest wall, Jean Corvisart des Marets was introducing the technique of percussion to Parisian physicians through his translation of Leopold Auenbrugger's 1761 treatise, Inventum novum ex percussione thoraces humani ut signo abstrusos interni pectoris morbos detegendi ("New Invention by Means of Percussing the Human Thorax for Detecting Signs of Obscure Disease of the Interior of the Chest"). The percussion technique was based on the empirical observation that striking a blow against a chest full of fluid would produce a dull sound. In contrast, if the chest was full of air, the blow would resonate for all to hear and for the percussor to feel.
In the absence of other techniques for thoracic examination and diagnosis, percussion was raised to a high art. Then, in their writings of the early nineteenth century, Pierre Piorry and Joseph Skoda advocated the use of two instruments for percussion: the plexor and the pleximeter. The plexor was a hammerlike device employed to strike a blow against the chest wall; the pleximeter was a flat, platelike object placed against the chest wall to give the examiner a uniform surface to strike. Eventually, it became dear that the same quality of examination was possible without those tools, and the two-handed method became standard. Moreover, the percussionist's vocabulary became standardized. By convention, five major types of sound are generated by percussion (see table).
One of the reasons percussion has lost favor in some quarters is that the conclusions of different physicians examining the same patient can vary greatly. This so-called interobserver variation is an important consideration in evaluating the worth of any test. Variation can result from the use of different percussive techniques: for example, one-handed versus two-handed or "firm" versus "light." It can result from differences in interpretation of findings: for example, from basing a diagnosis of splenomegaly on the spleen's position in relation to the midclavicular line, which often is mislocated. It can result from attaching different importance to the same finding. For example. one examiner might accord great significance to the detection of shifting dullness, while another might not.
In one study, two experienced physicians measured the diaphragmatic excursions in a series of 29 patients. Interobserver variability was high: The examiners agreed to within 2 cm only 60% of the time. In another study, nine physicians assessed patients' lung function on the basis of physical examination. There was marked variation on every parameter assessed by percussion: diaphragmatic excursions, pulmonary hyperresonance, and impaired cardiac and hepatic dullness.
The value of percussion is best appreciated in certain specific applications.
Finding and measuring the liver. In the 1960s, comparisons of physical diagnoses with radiologic and scintigraphic findings facilitated the development of a nomogram integrating the patient's sex. height, and weight in the assessment of liver size. The craniocaudad length of the dullness found on hepatic percussion could then be interpreted as "normal" or "enlarged." D. 0. Castell and colleagues, who described the original nomogram, noted that "soft" or "light" percussion tended to yield larger measurements than a "firm" percussive technique. A subsequent study suggested that light percussion led to measurements more closely resembling those obtained by abdominal ultrasound, perhaps because light percussing serves well for identifying the very thin lower edge of the liver. It was further suggested that for maximum accuracy in measurements, the light percussion should be done during full inspiration. Measurements obtained by firm percussion do not seem to vary significantly with respiration. In sum, firm percussion is more consistent than light percussion, but it regularly underestimates liver size as measured ultrasonographically.
Finding the spleen. Many techniques of spleen palpation have been described. They all have their drawbacks-but then, finding the spleen is one of the most unsatisfactory undertakings in physical diagnosis. Indeed, splenomegaly may be more difficult to identify than abnormalities of the jugular venous pulse or the subtle signs of constrictive pericarditis. For its edge to be palpable below the left costal margin, the spleen must be radiologically enlarged by 40% (One reason, evidently, is that initial enlargement of the organ is chiefly in an anterior direction.) Since a palpable spleen can be detected in less than 3% of the patients physicians routinely see, there is little opportunity for us to improve our skills as spleen detectives.
Percussion sometimes helps to locate a marginally enlarged spleen when no other maneuver is satisfactory. The patient must be examined in the supine position. The examiner percusses over the left eighth or ninth rib at the anterior axillary line (Figure 1, top). A resonant percussion note indicates normal spleen size. If, on the other hand, dullness is elicited or the quality of the sounds changes from resonance to dullness on full inspiration, the spleen will usually turn out (on ultrasonography or nuclear scan) to be enlarged.
Another way to identify splenomegaly is to percuss the stomach (Figure 1, bottom). Ordinarily, the stomach bubble will resonate. The field of resonance over the bubble is known as Traube's space. When Traube's space is to the right of its normal position, an enlarged spleen may have displaced it.
Finding pericardial effusions. Percussion of the heart is not a routine procedure-palpation usually yields the needed information. However, identification of the left border of cardiac dullness sometimes facilitates a diagnosis of pericardial effusion. A finding that the border is lateral to the point of maximum cardiac impulse may mean that the pericardium contains fluid.
In the late 1800s, William Ewart described a different method of identifying pericardial effusions. He recognized that when the pericardial sac contains fluid, the adjacent lung is compressed. The pericardial effusion can thus produce physical findings of lung consolidation, namely dullness to percussion , increased tactile fremitus, and egobronchophony. For that reason, Ewart's sign (an area of dullness at the angle of the left scapula) is indicative of pericardial fluid (Figure 2, left).
It should be kept in mind that massive cardiomegaly can produce similar lung compression and hence a similar constellation of findings. Identifying compression fir to pericardial effusion therefore requires careful percussion in the axilla and, posteriorly, at the angle of the scapula. Pleural effusion also compresses lung tissue. The primary area of percussive dullness may then be accompanied, beyond the involved lung base and past the spine, by Grocco's triangle, a paravertebral area of dullness over the normal lung (Figure 2, right). The base of the triangle extends 4 to 8 cm from the midspinal line. Physicians who know about Grocco's triangle may be able to spare their patients an unnecessary late-night thoracentesis.
Diagnosing ascites. Physicians know that percussing for ascites is difficult and unreliable. On the other hand, a percussion strategy called the puddle sign technique may permit identification of collections of fluid as small as 125 ml. The idea is to get a small amount of ascitic fluid to "puddle" in a place where it can be recognized. Accordingly, the patient is required to lie prone for five minutes and then get on all fours on the examining table. The diaphragm of the stethoscope is placed against the most dependent part of the abdomen, and the patient's flank is percussed with the middle finger, which is flicked briskly and repeatedly at a constant intensity against the abdominal wall. If the percussion note becomes more resonant as the stethoscope is moved up the flank, the physician has reason to think that fluid indeed has puddled in the dependent region. In their paper describing the puddle sign, J. D. Lawson and A. S. Weissbein argue that the sign is more reliable than other physical signs of ascites because it measures sound transmission, whereas the traditional percussion tests for shifting dullness and flank dullness measure sound absorption.
When abdominal ultrasonography was first introduced, the detection of ascites was used to compare the new technique and physical diagnosis. In one study, the findings of three experienced examiners were compared with each other and with ultrasonographic findings. The sensitivity of the physical examination turned out to range from 50% to 94%, and the specificity from 29% to 82%. The absence of flank dullness was the only reliable physical sign: It permitted ascites to be excluded with an accuracy of 90%, The puddle sign was no better than any other percussion test. However, one examiner was much more accurate than the other two in the use of the sign, which confirms that a particular maneuver may be valuable in the hands of a skilled examiner. The study also may underscore the problem that interobserver variability poses even among the most experienced clinicians.
Processes involving the lung apex. The apices of the lungs are difficult to visualize by routine chest roentgenography. Here percussion can be useful. The examiner's intent is to assess the width of Kronig's isthmus, a band of resonance between the neck and the shoulder girdle muscles (Figure 3). The Isthmus is narrowed by fibrotic contraction or consolidation at the lung apex. In obese patients, Kronig's isthmus may not be identifiable.
Plexors and pleximeters are not in wide use, but the basic two-handed percussion technique is sometimes modified in other ways.
Auscultatory percussion employs the stethoscope to assess sound transmission through the thorax. Its proponents (like those who champion the puddle sign) argue that the technique is much more sensitive than standard percussion methods. It may indeed be superior to simple percussion for identifying lesions deep in the lung parenchyma; simple percussion loses value in detection of lesions more than 5 cm inward from the skin. In our clinic it has produced dramatic findings: We once located lung nodule by means of auscultatory percussion. Lesions smaller than that have been identified, sometimes when the chest roentgenogram and CT scan were unrevealing. Auscultatory percussion can also be useful as a screening test: An abnormal sound from a particular lung zone may prompt the physician to obtain a chest film.
To perform auscultatory percussion, the examiner places the stethoscope over the posterior thorax and percusses the sternum with the middle finger. Each lung zone is compared with its contralateral counterpart; the presence of a solid lung lesion will increase sound transmission, so that the amplitude of the percussion note will be increased on that side. If fluid is present, sound transmission will be reduced. The paravertebral areas are examined in a similar manner in an effort to identify mediastinal masses.
The limiting factor in the use of auscultatory percussion is that the patient must have no underlying pulmonary pathology. A patient with bronchitis and emphysema may give extremely abnormal findings due to lung densities at various points of examination, particularly If there are endobronchial secretions, yet the chest film will reveal nothing but the usual manifestations of chronic obstructive pulmonary disease.
One-handed percussion was described by Auenbrugger in his treatise of 1761. In essence, the examiner uses a plexor without a pleximeter. (One hand strikes, but the other does not receive the blow.) The technique is now considered inferior to two-handed percussion. Nevertheless, one-handed percussion offers a way to accentuate positive findings. Pneumothorax is an example. When there is massive collapse of a lung, the hemithorax will be tympanic: It will sound like a drum when it is struck with the fingers of one hand. Such a finding should prompt an order for a chest film and emplacement of a chest tube, not necessarily in that order.
Multifinger percussion. Physicians sometimes place three or four fingers on the chest wall and strike each of them in turn. The technique is advantageous when it helps the examiner ascertain where the percussion note changes. It is disadvantageous in that keeping several fingers on the chest wall dampens sound transmission and hence introduces a major source of error. In general, multifinger percussion is difficult to execute reliably.
Percussion is part of the history and physical examination. Nevertheless, we should think of it much as we think of any radiographic, biochemical, or hematologic study done as part of the diagnostic process. We cannot know what a study means unless we know how it was done and under what conditions. The same is true of percussion. There are many ways to percuss a chest or abdomen - perhaps as many as are physicians to do the percussing. But in order for each of us to draw the right conclusions from our findings, we must understand the characteristics and limitations of our individual techniques. Above all, we need to do our percussion the same way every time, or we will never get a sense of normal and abnormal findings.
Percussion can then claim its place as a useful adjunct to other methods of examination-or as an indispensable method when more sophisticated tests are unavailable or when the findings are so critical that we trust our own senses best. Knowing more about this old skill can make our daily rounds more interesting and can lead to unexpected discoveries. For their part, the modern diagnostic technologies offer a good standard against which we can measure the relative value of quick and "handy" traditional methods, such as percussion, to help us decide when we need to exceed their sensitivity and specificity as we "lay hands" on our patients.
Hospital Practice, September 30, 1987, pp 25-36