auscultation of the lung: past lessons, future possibilities · auscultation ojthe lung:past...

Thorax, 1981, 36, 99-107

Auscultation of the lung: past lessons, futurepossibilitiesRAYMOND L MURPHY

From the Pulmonary Service, Lemuel Shlattuck and Faulkner Hospitals,Boston, Massachusetts, USA

ABSTRACT Review of the history of auscultation of the lung reveals few scientific investigations. Themajority of these have led to inconclusive results. The mechanism of production of normal breathsounds remains uncertain. Hypotheses for the generation of adventitious sounds are unproven.Advances in instrumentation for lung sound recording and analysis have provided little of clinicalvalue. There has been a recent resurgence of interest in lung sounds. Space-age technology has im-proved methodology for sonic analysis significantly. Lung sounds are complex signals that probablyreflect regional lung pathophysiology. If they were understood more clearly important non-invasivediagnostic tools could be devised and the value of clinical auscultation could be improved. A multi-disciplinary effort will be required to achieve this.

"Every adversity has within it the seed of anequivalent or even greater benefit." (N Hill)Laennec revolutionised the diagnosis of lung diseasewith an instrument that is imprecise by today'sstandards. Clues to the explanation for this surprisingfinding can be found in the following review of lungsound literature. Although much has been writtenon lung sounds, few scientific investigations havebeen reported. In this article I will focus on studiesof the mechanism, production, and transmission ofnormal and abnormal lung sounds as well as methodsfor objective analysis of these sounds. These scientificstudies provide interesting lessons and inspirespeculation on the potential value of a more completeunderstanding of pulmonary auscultation.

Origin of vesicular and bronchial sounds

Laennec' described the normal lung sound as "aslight but extremely distinct murmur answering tothe entrance of the air into and its expulsion fromthe air cells of the lungs." Early speculation includedthe following: (1) the noise was made all along therespiratory tract by friction of gases against theairway walls; (2) the glottis alone caused the noise-the sound heard over the chest wall was the glotticsound attenuated during transmission through the

Address for reprint requests: Dr RLH Murphy, Director, PulmonaryService, Faulkner Hospital, 1153 Centre Street, Boston, MA 02130,USA.

lung; (3) the sound was made as air passes from anarrower to a wider space on respiration. Bullar2tested these hypotheses by using an artificial thoraxconsisting of an air-tight, fluid-filled chamber withglass sides. He could expand and contract thischamber at will with a bellows. He placed one oftwo sheep lungs connected to each other by thetrachea and bronchi into the artificial thorax, theother being left outside. With the trachea open,operation of the bellows ventilated the lung in thebox while the outer lung lay still and collapsed.Vesicular sounds were heard over the lung in the box;bronchial sounds were noted over the outer lung. Thewell-recognised phenomenon of sound productionwhen an air current passes over but not into themouth of a tube was considered the mechanism forthe bronchial breathing in the outer lung. The tracheawas then plugged so that during ventilation of theinner lung the air was displaced in and out of theouter lung. Vesicular sounds were then heard overthe outer lung indicating that they were not producedin the glottis or trachea alone. Bullar also demon-strated that no sound was generated despite fluctua-tions in lung volume and pressure under conditionsof no flow. He concluded that sounds are producedat those parts of the respiratory tract where air passesfrom a narrower to a wider space. Thus, duringinspiration one sound is produced at the glottis andanother at the points where the bronchioles openinto "the vesicles."

99

copyright. on S

eptember 10, 2019 by guest. P

rotected byhttp://thorax.bm

j.com/

Thorax: first published as 10.1136/thx.36.2.99 on 1 F

ebruary 1981. Dow

nloaded from

http://thorax.bmj.com/

RaymondL Murphy

Hannon et a13 investigated the site of lung soundproduction by placing exenterated sheep lung overthe neck of a human and recording the frequencycharacteristics of the sounds heard over the sheeplung and over the back in the tenth intercostal space.They found no vibrations below 130 Hz over thesheep's lung but such vibrations were simultaneouslydetectable at the tenth intercostal space. Theyconcluded that ventilation of lung underlying thestethoscope was necessary to produce this lowfrequency soundAs sound filtering techniques were developed

they began to be applied to lung sound analysis.German investigators tried to relate the frequentcharacteristics of bronchial breathing to the size ofthe airway in which they were produced.4 The under-lying assumption was that the various portions of thetracheobronchial tree vibrate on excitation atfundamental frequencies related to their dimensions.They believed that such knowledge would provideinformation on the location and extent of consoli-dated lung tissue. Bronchi or the model rubber tubesthey studied, which had internal diameters of 3 mmor less, were poor conductors of sound of the fre-quency of bronchial breath sounds. Since they wereunable to record vibrations of 1700 or greater inbronchial breath sounds they concluded that airways3 mm in diameter or less take no part in the formationof such sounds. They concluded that consolidatedlung tissue had to extend from the pleural surfaceto the conducting airways of 2-3 mm in size to hearthe bronchial sounds. Their results correlated withmeasurements in cadaver lungs and clinical findingsin pneumonia.

Increased information on pulmonary function,especially the mechanics of breathing, led to attemptsto use this knowledge in interpreting lung soundproduction. Forgacs5 believed that the source ofthe lung sounds is the turbulence in the airstreamflowing through the pharynx, glottis, and largeairways. The extent of the turbulence peripherallydepends on the flow rate, but even in the restingsubject there is often turbulence in the main bronchi.He called attention to the clinical observation thatbreath sounds in some subjects are silenced syn-chronously with the heartbeat while they are heardcontinuously on the opposite side. He deduced fromthis that part of the noise must be generated at leastas far peripherally as the main bronchi and possiblyin more peripheral branches as well. This componentof the breath sounds is suppressed when the velocityof the airstream entering the left lung is momentarilyslowed by air driven in the opposite direction bypulsation of the heart. Forgacs6 quantified theincrease in sound at the mouth in bronchitis andasthma and the absence of such an increase in

primary emphysema. This also argued for an infra-glottic source of the inspiratory sound especially asthe increased intensity of the respiratory sound wasalso present over the stoma of a patient with bron-chitis who had undergone laryngectomy. When ahelium mixture was breathed by these patients theinspiratory sound became quieter, consistent withthe theory that turbulence was the cause ofthe sound.Forgacs argued that calculations from the rate ofvolume flow at the mouth and the total cross-sectionof the airways at each generation of the bronchiindicated that the flow velocity beyond the lobaror segmental bronchi is too low to generateturbulence. In the more peripheral branches flowwas laminar and presumably silent.Hardin and Patterson7 applied flow visualisation

techniques to the study of the behaviour of air in amodel of a typical junction in the human bronchialtree. Vortices were formed at each junction atReynold's numbers in the range 50-4500 correspond-ing to those found in the fifth through thirteenthgenerations of the bronchial tree. Combining thiswith the theory of vortex sound generation, theyconcluded that the frequencies of sound producedby these vortices corresponded to those seen onspectral analysis of lung sounds. Furthermore, theystated that the frequency of sound produced at eachvortex was related to the flow velocity and diameterof the tube. Thus, they believed that the order ofbronchi in which particular frequencies are generatedcan be determined. Unfortunately, the proof thatthe vortices themselves produce the sound or thatthe spectral components of the lung sounds theyanalysed came from the size bronchi they claim isnot convincing.To study the site of production of breath sounds,

Kraman8 made simultaneous recordings from twodifferent areas of the chest. These were then analysedby subtraction pneumography to determine whetherthey were similar sounds, and therefore, transmittedfrom a distant source or dissimilar, and therefore,transmitted from sources located near each micro-phone. His findings using microphones separatedby 10 cm were consistent with an intrapulmonarysource for the inspiratory component and an upperairway source for at least some of the expiratorycomponent. This work represents a clever applicationof a standard acoustic technique to help identifythe site of lung sound production.

Curiously, investigators who actually performedexperiments to observe sounds concluded that somesound is generated peripherally despite the theo-retical arguments that suggest that air movement inthe periphery is silent. Were the answer known itwould allow a more precise estimate of how usefullung sounds are in estimating regional ventilation.

100

copyright. on S



j.com/


ebruary 1981. Dow

nloaded from


Auscultation oJ the lung: past lessons, future possibilities

A discussion of this topic, however, requires con-sideration of the sound conducting properties of thelung.

Transmission of sounds in the chest

Bullar2 had observed in his model that bronchialsounds were attenuated when the inner lung wasdistended and louder when it was collapsed, con-sistent with the clinical observation that bronchialsounds are attenuated by air filled lung. The graphicrepresentations of respiratory sounds of Hannon andLyman3 were made (1) directly over the trachea, (2)transmitted through a constant thickness of sheep'slung placed over the neck of the same subject, and(3) at the right base of the chest of the same subject.Interestingly, the sound they observed directly overthe trachea in expiration without filters was louderthan the inspiratory sound but the inspiratory soundbelow 130 Hz was louder than the expiratory soundbelow 130 Hz.Using a low pass filter to exclude vibrations

greater than 130 Hz they observed low frequencyvibrations over the trachea in a normal human.These were also present when an uninflated sheep'slung was interposed between the recording electricalstethoscope and the larynx of this subject. When thelung was inflated to a thickness of 4 cm these vibra-tions were not recorded no matter what the pressurein the lung. The low frequencies were visible onsimultaneous recordings made at the right lung base.They concluded that the passage of air into and outof the lung was necessary for the transmission ofthese low frequency vibrations. Nairn and Turner-Warwick9 in 1968 compared breath sound intensityto regional ventilation estimated by radioactivexenon. When sounds were poor or absent in emphy-sema there was almost invariably poor intrapulmon-ary mixing in the underlying lung zone. This isconsistent with an intrapulmonary origin ofthe soundbut does not exclude a central origin with regionalalterations in transmission characteristics. LeBlanc,Macklem, and Ross'0 studied breath sound intensityunder controlled lung volumes, flows, and bodypositions. In subjects in the upright and lateraldecubitus positions they found that maximal relativeintensity of breath sounds varied in the same way asexpected from studies of the regional distributionof ventilation observed with radioactive xenon. Theybelieved that breath sound intensity was a goodphysical sign of regional pulmonary ventilation andpresented evidence that the vesicular sound wasrelated to alveolar filling. In contrast, Pioy-Song-Sang et all" found poor correlation with a phono-pneumographic breath sound index and regionaldistribution of xenon and concluded the breath

sound heard by clinicians cannot be used to assessregional ventilation. They corrected the breathsound index by dividing it by an index of soundtransmission obtained by introducing white noiseat the mouth. The breath sounds so "conioensated"were believed to reflect not only regional ventilationbut also the number of sound generators under thenicrophone.Rice12 pointed out that sound pronagates along

the airways, with relatively little attenuation, at freefield sound speeds in the gas, but that sound speedsin the lung were reported to be unusually slow. Thissuggested that the low velocities were caused by apropagation along a path much longer than thestraight line distance between the sound source andreceiver. He measured velocities ranging from 20 to70 metres per second across excised horse lung.This was less than the known speed of sound in air(300 m/s) or solid tissue (1500 m/s). Inflation of thelung with high or low density gas changed thevelocity only slightly, demonstrating that circuitousroutes of sound propagation through airways wereunlikely. The sound velocities he observed approxi-mated those calculated assuming that the parenchy-ma is a uniform fluid and that compressibilityand density are the average of the gas and tissuephases. This is the most enlightening work on thesubject to date, but awaits confirmation in humans.The relevance of transmission of sound across thelung to sound generated within it should also beclarified.The study of lung sound generation and trans-

mission in disease states is even more difficult thanin normals because of the occurrence of adventitiousor "added" sounds. The understanding of theselatter sounds is further complicated by the problemoflung sound nomeclature.

Nomenclature ani observer variability

Laennec said lung sounds were easier to distinguishthan describe. The difficulties in description have ledboth to a plethora of terms for the same sound andto distinctly different sounds being called by a singlename. The confusion caused by the translation ofL'Auscultation Mediate into English was describedby Robertson'3 in an elegant, albeit whimsical review.Confusion in terms is probably an important factorin observer variability. Perceptual differences betweenobservers may also exist and the sounds themselvesvary from breath to breath. In contrast, heart soundsare relatively uniform from beat to beat, perhapsaccounting for their greater diagnostic utility. Therehave been relatively few studies to quantify themagnitude of observer variability. Most of thesehave focused on reliability or observer agreement

101

copyright. on S



j.com/


ebruary 1981. Dow

nloaded from


RaymondL Murphy

rather than accuracy. Most studies have shownhigh disagreement,14-'8 a problem in clinical datacollection in general,18 including the interpretationof chest radiographs.19 20 Some studies9 21 haveshown fairly good agreement, perhaps by employingmore rigid criteria and methodology or by focusingon single observations at a time. The accuracy oflung sounds in detecting obstructive lung diseaseappears to be better than other physical signs.2' 22Accordingly, objective methods to quantify lungsounds have been sought.

Lung sound recording and analysis

Numerous attempts have been made to displaysound visually. They have focussed on the amplitudeof the sound over time, or temporal properties, andon analysis of the sound by component frequencies,or spectral properties.23 35 In 1924 Bass36 reporteda clever method using a condenser microphone andoscilloscope that displayed normal and abnormallung sounds. Cabot and Dodge3' were able to showin 1925 that coarse and fine rales were associatedwithconspicuous low and high frequency components

respectively. In 1955, McKusick et a123 demonstratedthat the spectrograms of "moist "and "dry "cracklingrales were different, the dry rales containing morehigh frequencies. Calibrated amplitude plots re-ported by Weiss and Carlson24 in 1970 allowedvisual presentation ofthe overall amplitude ofsounds,including pauses and duration differences as well as astudy of the relation of the amplitudes in inspirationand expiration. Adventitious sounds were notclearly distinguished from one another by thesemethods nor by the integrating envelope detectordescribed by Wooten and Waring.27

Speculations on the difficulty of visually differenti-ating lung sounds were presented by Forgacs in1967.37 He pointed out that the study ofcrackling wasdifficult because the sounds follow one another sorapidly that neither the individual crackles, nor therhythmical pattern made by their sequence can beidentified with the unaided ear. He noted that theoscilloscope showed the regular waveforms of awheeze but only at time base speeds at which onesaw such a small section of sound that it was verydifficult to judge the timing of the wheezes in relationto the respiratory cycle.

Table Outline ofclassification oflung sounds.

Acoustic characteristics Waveform Recommended Terms in A British Laennec's Laennec's modelATS* some usage original termnomenclature textbooks

Discontinuous, C C Coarse crackle Coarse rale Crackle Rale muquex Escape of waterinterrupted explosive ou gargouille- from a bottle heldsounds -A %A %,/_ ment with mouthLoud, low in pitch directly downward

Discontinuous, C Fine crackle Fine rale Crackle Rale humide Crepitation ofinterrupted explosive C crepitation ou crepitation salts in a heatedsounds dish. NoiseLess loud than above emitted byand of shorter healthy lungduration; higher in when compressedpitch than coarse rates in the handor crackles

Continuous sounds Wheeze Sibilant High-pitched Rale sibilant ProlongedLonger than 250 ms, rhonchus wheeze sec ou whisper of varioushigh-pitched; dominant | sifflement intonations;frequency of400 Hz or V Y W chirping of birds;more, a hissing sound sound emitted by

suddenlyseparating 2portions ofsmooth oiledstone. Themotion of a smallvalve

Continuous sounds A A Rhonchus Sonorous Low-pitched Rale sec Snoring; bassLonger than 250 ms rhonchus wheeze senore ou note of a musicallow pitched; dominant ronflement instrument;frequency about 200 Hz cooing of a woodor less; a snoring sound pigeon

*American Thoracic SocietyWhile the terms used to name the categories oflung sounds vary widely, the categorisation scheme itselfhas changed little since Laennec. The mostrecent names recommended for adoption by the American Thoracic Society and terms used by others are shown here, accompanied by acousticdescriptions and examples of typical sound waveforms for each category.

102

copyright. on S



j.com/


ebruary 1981. Dow

nloaded from


Auscultation of the lung: past lessons, future possibilities

Forgacs,37 38 and Nath and Capel39 40 describedthe repetitive nature of rales in some patients buttheir published tracings had time scales below100 mm/s and did not reveal clearly the waveformcharacteristics of the adventitious sounds. Theamplitude deflections produced by rales, for example,are similar to deflections found in tracings in whch norales are described since the specific waveforms ofthe rales are obscured at this speed.

Space-age technology has provided more sophisti-cated computer-based methods, such as time-A NORMAL TRACHEAL

expanded waveform analysis, which have helped tosolve some of the nomenclature problems (figure).Focusing on the objective description provided byauthors as well as using time-expanded waveformanalysis for verification has enabled the developmentof the classification outlined in the table. Ladnnec,'with his characteristic thoroughness, provided uswith a variety of models. Time-expanded waveformanalysis of sounds made by these models shows thatthey are similar to those of the lung sounds they wereintended to represent. For example, sound produced

C FINE CRACKLES

C

- c~~

B NORMAL VESICULAR D WHEEZE

XCM >VI iiMWR AR ~~~~~~~~~~~~VU v/N-APJW_eV

'AiWVAWVurW_'1'-/\Ap-

120 ms

Figure Time-expanded waveforms oflung soundsThe time-intensity plots in this figure are madefrom tape recordings oflung sounds stored in a computer memory andthen visually displayed on a two dimensional plot, with amplitude or intensity on the Y axis. The time axis of the plot is"expanded" or magnified by playing back slowlyfrom the computer memory.Plots A andB illustrate normal trachealand vesicular sounds. The louder, longer expiratory phase ofthe tracheal sound is readily recognised as is the pausebetween inspiration and expiration. When crackles occur, they produce intermittent ("discontinuous") deflectionssuperimposed on the normal vesicular pattern (marked C in plot C). Plot shows a "continuous" deflection produced by awheeze replacing the normal waveform.

q4j*~

r}

103

copyright. on S



j.com/


ebruary 1981. Dow

nloaded from


104

by heating salt in a frying pan shows intermittentspikes similar to those offine crackles, while a musicalnote shows undulating sinusoidal deflections similarto a rhonchus. More rigorous analysis of lung soundscoupled with establishment of standards for theclassical major categories should allow even furtherclarification of the nomenclature.

DISCONTINUOUS SOUNDS

Very few people have attempted to define either thesites or origin or the mechanism of production ofadded or adventitious sounds. The most clearlythought-out efforts in this regard are those ofForgacs.41 It seems likely that more than one generat-ing mechanism for crackles exists. The coarse,

discontinuous sound occurring in pulmonary oedemaor other conditions in which fluid is found in theairways are probably caused by rupture of fluid filmsor bubbles. Forgacs37 has argued convincingly thatcrackles occurring in diseases such as pulmonaryfibrosis or early congestive heart failure are not aresult of air bubbling through fluid. His observationssupport instead the hypothesis that crackles in thesediseases result from the sudden release of energystored in the lung after delayed opening duringinspiration of airways that had closed at the end ofthe previous expiration. He based his arguments onthe often strikingly repetitive occurrence of cracklesduring succeeding inspirations, even after a cough.These crackles rarely occur in expiration and are

often localised over dependent areas of the lung,where gravitational stress predisposes the airwaysto collapse at low lung volumes. Air bubbling throughfluid would be expected to cause a noise during bothinspiration and expiration. The pattern would pro-

bably change after a cough had redistributed the fluid.

CONTINUOUS SOUNDSWheezes are believed to be generated by a regularvibration, or oscillation, of the airway wall at one ormore sites in the chest. Forgacs41 observed that thepitch of a wheeze is relatively independent of gasdensity. This indicates that wheezes are not producedin the same way as the note of an organ pipe, in whichthe pitch of the note depends both on the length ofthe pipe and on the density of the vibrating gas in thepipe. Forgacs also found that he could generatewheeze-like noises in excised bronchus by blowingair through the specimen and using compression togenerate external pressure.

Thus, it seems likely that wheezes occur when airpassing through a narrowed airway at high velocityproduces a decrease in the gas pressure in the airwayat the region of constriction. Bernoulli's principlerelates the local velocity of the gas to local pressurein such flow. As the velocity of the gas increases at

RaymondL Murphy

the constriction, the pressure decreases. If allowedby the other forces acting on the airway, collapse willcontinue progressively until there is a substantialresistance and the flow is decreased. Then the internalpressure increases and the lumen enlarges. Thisalteralion of the wall of the airway between almostclosed and almost open is believed to produce thewheezing noise.Our understanding of adventitious sounds is even

more incomplete than our knowledge about thenormal sounds. Like the latter it is based on too fewexperiments and much speculation. The numerousstudies that have attempted clinical correlation oflung sounds with disease states have been deficient.Most have had too few subjects, outdated equipment,or more importantly, conclusions that were imprecise.It is not surprising that an important text of pul-monary disease had the following comment: "Thestethoscope is largely a decorative instrumentin so far as its value in diagnosis ofpulmonary diseasesis concerned. Nevertheless, it occupies an importantplace in the art of medicine. Apprehensive patientswith functional complaints are often relieved assoon as they feel the chest piece on their pectoralmuscles." 42

Current clinical uses

It seems paradoxical that despite our incompleteinformation on lung auscultation and the availabilityof more accurate diagnostic tools, physicianscontinue to use the stethoscope routinely to examinethe lungs. It is useful to consider the reasons for this.Lung sounds can be quite important in the dailybedside and outpatient management of patientswith a variety of illnesses such as bronchial asthmaand congestive heart failure. These sounds canprovide crucial information in instances of upperairway obstruction, misplaced endotracheal tubes,and spontaneous pneumothorax. In contrast to manydiagnostic techniques, auscultation is readily per-formed in a variety of settings. It is often used to helpdecide on the necessity for additional clinical studies.It is my belief that more precise knowledge wouldimprove its accuracy in such instances. More impor-tantly, understanding the nature of lung soundscould lead to improved non-invasive pulmonarydiagnosis. The reasons for this belief will be con-sidered now.

Areas for further investigation including potentialfuture applications

An extremely complex sonic signal is sent by the lungduring breathing. Modern analytic techniquesprovide the equivalent of an acoustic microscope.

copyright. on S



j.com/


ebruary 1981. Dow

nloaded from


Auscultation of the lung: past lessons, future possibilities

This allows study of the lung's sonic signal with aprecision not previously available. This signal can beseen to vary with recording site, flow rate, lungvolume, body position, and various breathingmanoeuvres. It is likely that the sound changes withgrowth, development, and age, as well as withminimal environmental insults. The signal is socomplex and varies so much that it appears at timesto be random or unpredictable. It is more likely,however, that the sonic signal reflects the underlyinganatomy and pathophysiology. Were it better under-stood, powerful diagnostic tools using lung soundmaps might be developed. In contrast to pulmonaryfunction studies, the information obtained is likelyto be determined by regional pathophysiology.Lung sounds promise to reflect certain pathologicalprocesses such as the obstructive diseases and earlyinterstitial processes more closely than radiologicalstudies do. The credibility of such speculation isenhanced by considering the advances that came inchest diagnosis from a simple wooden tool. Much ofthis success is attributable to the remarkable dis-criminatory powers of the human ear-brain systemand to those who made objective studies. It is alsolikely that this came about because of the wealth ofthe information contained in this signal.There are numerous problems. The major problem

is that the potential amount of information is toogreat to be processed readily with currently availableinstrumentation. Lung sounds have not beenmodelled adequately. This limits predictions as tothe factors which affect them. Their diagnosticutility may also be limited if differing pathologicalprocesses produce similar sounds. Indeed, thecrackles of congestive heart failure are believed bymany to sound similar to those of interstitial fibrosis.Recent observations using advanced analyticaltechniques suggest that they contain measurabledifferences.43 The extent to which such observationswill be fruitful is not yet known. However, precisecounting of crackles, their measurement in terms oftheir individual characteristics, and study of theirrelative timing in the respiratory cycle is in its infancy.It is likely that noninvasive techniques will allowprecise location of the origin of individual crackleswithin the lung using such techniques as triangulation,developed to detect submarine location. This willprovide a means of clearly identifying minutechanges in pulmonary pathology within a variabletime frame, for example, from minute to minute inintensive care or even from year to year as in monitor-ing of industrial lung disease. There are many knownbut quite curious clinical facts that if properlyquantified could allow more precision in noninvasivediagnosis. For example, the crackles caused byasbestosis are numerous, while the number of such

crackles which occur in sarcoidosis with radiographicchanges of equivalent severity are relatively few.Knowing the precise quantity and distribution overthe chest wall of such sounds is likely to be useful indiagnosis. It might be argued that such informationwill require advanced technology which will be tooexpensive to be practical. This may be true but it isprobable that much of the complex instrumentationrequired to solve the problems outlined above willnot be necessary in transposing the advances inknowledge to practical use. An analogy to illustratethis is the demonstration that timing in cracklesrelates to the presence of bronchitis or obstructivelung disease; early crackles indicate bronchitis, lateindicate interstitial fibrosis.40 Once the nature of thelung sound is more precisely known, relativelysimple means could be developed to study specificaspects which have been shown to be clinicallyrelevant. If the exact variation in frequency contentof sound at a specific chest wall site were known innormals, simple devices to detect abnormalitiesmight be developed to screen workers in industrialdisease surveys. Electronic processing of soundsignals has become greatly simplified with theadvances in microcircuitry and the decrease inexpense of computer memory. It is unrealistic tothink that such devices will not be even more readilyavailable in future years than they currently are.

In summary, if the huge amount of informationcontained in the lung sounds could be interpretedthe potential exists for very powerful diagnosticinstruments. To understand lung sounds clearlywill require the efforts of a variety of disciplinesincluding anatomy, physiology, pathology, acoustics,mechanical and electrical engineering, psycho-acoustics, epidemiology, statistics and computerscience. The latter science has tremendous advantagesin information processing not possessed by the ear-brain system. An interdisciplinary effort has alreadybegun in the form of the International Lung SoundsAssociation which will have its sixth annual meetingin October. Much of the work presented at the firstfive meetings has not yet been widely publicised.The problems with interdisciplinary approachesappear to become more complex as each disciplineadvances. This was not apparently the problem forLaennec who auscultated patients and performednecropsies. It is almost embarrassing, but indeed atribute to him, that his work stands now in 1981 asthe most comprehensive in correlation of pathologywith clinical lung auscultation. His example ofcareful observation and correlation needs to befollowed to turn the adversity of "too much" in-formation to diagnostic advantage.

I would like to thank the following p3ople for their

105

copyright. on S



j.com/


ebruary 1981. Dow

nloaded from


106

assistance: Robert G Loudon, Sadamu Ishikawa,Peter Workum, Frank Davidson, Hans Lipp,Stephen Holford, Elizabeth Del Bono, and Lee Starr.The work was supported partly by grants from theInstitute of Occupational and Environmental Health,Montreal, PQ, Canada, and by the US NationalHeart, Lung, and Blood Institute (no HL23318).The table is reproduced with permission from ASimplified Introduction to Lung Sounds by R L HMurphy, published by Stethophonics, PO Box 122,Wellesley Hills, MA02181, USA.

References

1 Laennec RTH. A treatise on the diseases of the chestand mediate auscultation. Translated from the Frenchedition by John Forbes. New York: Samuel Woodand Sons, 1935.

2 Bullar JF. Experiments to determine the origin of therespiratory sounds. Proc R Soc London 1884; 37:41 1-23.

3 Hannon RR, Lyman RS. Studies on pulmonaryacoustics. II. The transmission of tracheal soundsthrough freshly exenterated sheep's lungs. Am RevTurberc 1929; 19:360-75.

4 Martini P, Muller H. Studies on bronchial breathing.Deutsche Arch FKlin Med 1923; 143:159-73

5 Forgacs P. Lung sounds. Br J Dis Chest 1969;63:1-12.

6 Forgacs P, Nath AR, Richardson HD. Breathsounds. Thorax 1971; 26:288-95.

7 Hardin JC, Patterson JL Jr. Monitoring the state ofthe human airways by analysis of respiratory sound.Acta Astronautica 1979; 6:1137-51.

8 Kraman SS. Determination of the site of productionof respiratory sounds by subtraction phonopneumo-graphy. Am Rev Respir Dis 1980; 122:303-9.

9 Nairn JR, Turner-Warwick M. Breath sounds inemphysema. BrJ Dis Chest 1969; 63 :29-37.

10 LeBlanc P, Macklem PT, Ross WRD. Breath soundsand distribution of pulmonary ventilation. Am RevRespir Dis 1970; 102:10-16.

11 Ploy-Song-Sang Y, Martin RR, Ross WRD, LoudonRG, Macklem P. Breath sounds and regionalventilation. Am Rev Respir Dis 1977; 116:187-99.

12 Rice D. Sound speed in the parenchyma of excisedhorse lungs. Presented at the Fifth InternationalLung Sounds Conference, London, England, 1980.

13 Robertson AJ. Rales, rhonchi, and Laennec. Lancet1957; 1:417-23.

14 Schilling RSF, Hughes JPW, Dingwall-Fordyce.Disagreement between observers in an epidemio-logical study of repiratory disease. Br Med J 1955;1:65-8.

15 Fletcher CM. The clinical diagnosis of pulmonaryemphysema. Proc R Soc Med 1952; 45:577-84.

16 Schneider IC, Anderson AE. Correlation of clinicalsigns with ventilatory function in obstructive lungdisease. Ann Intern Med 1965; 62:477-85.

17 Hudson LD, Conn RD, Matsubara RS, Pribble AH.

RaymondL Murphy

Rales-diagnostic uselessness of qualitative ad-jectives. Am Rev Respir Dis 1978; 113:187.

18 Koran LM. The reliability of clinical methods, data,and judgements. NEnglJMed 1975; 293:642-701.

19 Felson B, Morgan WKC, Bristol LJ, Pendergrass EP,Freesen EL. Observations on the results of multiplereadings of chest films in coal miners' pneumocon-iosis. Lintonow's Reger RB Radiol 1973; 109:19-23.

20 Cochrane AL. Observer error in the interpretation ofchest films. Lancet 1952; 2:505-9.

21 Pardee NE, Martin CJ, Morgan EH. A test of thepractical value of estimating breath sound intensity.Chest 1976; 70:341-4.

22 Schneider IC, Anderson AE. Correlation of clinicalsigns with ventilatory function in obstructive lungdisease. Ann Intern Med 1965; 62:477-85.

23 McKusick VA, Jenkins JT, Webb GN. The acousticbasis of the chest examination: studies by means ofsound spectrography. Am Rev Tuberc 1955; 72:12-34.

24 Weiss EB, Carlson CJ. Recording of breath sounds.Am Rev Respir Dis 1972; 105:835-9.

25 Grassi C, Marinone G, Morandini C, Pernice A,Puglisi M. Normal and pathological respiratorysounds analyzed by means of a new phonopneumo-graphic apparatus. Respiration 1976; 33 :315-24.

26 Cugell DW. Use of tape recordings of respiratorysound and breathing pattern for instruction inpulmonary auscultation. Am Rev Respir Dis 1971;104:948-50.

27 Wooten FT, Waring WW. Spectral and intensityanalysis of respiratory sounds. Proceedings of the25th annual Conference on Engineering in Medicineand Biology. Arlington, Virginia: Alliance forEngineering in Medicine and Biology, 1972.

28 Ebina T, Toshima M, Fukushi K et al. KlinischeAnwendung elektrisch registrierter Atemgerausche.Beitr Klin Tub 1956; 115:276-81.

29 Klebanov MA. Ronorespirograflia i perspectivy eeispol'zovaniia. Ter Arkh 1969; 41:57-63.

30 Zizlin DM, Rozenblat VV, Likhacheva El. Ob'ektiv-noe issledovanie dykhatel'nykh shumov s pomoschu-'iu chastotnogo analiza. Ter Arkh 1969; 41:108-13.

31 Cabot RC, Dodge HF. Frequency characteristics ofheart and lung sounds. JAMA 1925; 84:1793-5.

32 McKusick VA. Cardiovascular sound in health anddisease. Baltimore: Williams and Wilkins, 1958.

33 Murphy RLH Jr, Sorensen K. Chest auscultation inthe diagnosis of pulmonary asbestosis. J Occup Med1973; 15:272-6

34 Murphy RLH Jr. Human factors in chest ausculta-tion. In: Picket RM, Triggs TJ, eds. Humanfactors inhealth care. Lexington, Massachusetts: DC Heath,1975.

35 Jalil A, Aronovitch M. Recording and analysis ofbreath sounds in patients of chest disease. Indian JChest Dis 1966; 8:158-94.

36 Bass E. Objective recording of sound patterns ofnormal and pathological breath sounds. Ges InnMed 1927: 39:345.

37 Forgacs P. Crackles and wheezes. Lancet 1967;2:203-5.

copyright. on S



j.com/


ebruary 1981. Dow

nloaded from


Auscultation of the lung: past lessons, Juture possibilities 107

38 ForgacsP.Lungsounds.BrJDisChestl969;63:1-12. 41 Forgacs P. Lung sounds.. London: Bailliere Tindall,39 Nath AR, Capel LH. Inspiratory crackles and 1978.

mechanical events of breathing. Thorax 1974; 42 Rubin EH, Rubin M. Thoracic diseases. Philadelphia29:695-8. and London: WB Saunders, 1961.

43 Holford S. Lung sound analysis in the diagnosis of40 Nath AR, Capel LH. Inspiratory crackles-early and pulmonary asbestosis. SM thesis, MIT, Cambridge,

late. Thorax 1974; 29:223-7. Massachusetts, USA, 1976.

copyright. on S



j.com/


ebruary 1981. Dow

nloaded from


auscultation of the lung: past lessons, future possibilities · auscultation ojthe lung:past...

Documents