CLINICAL, RADIOGRAPHIC, AND PATHOLOGICAL STUDY OF ... · A CLINICAL, RADIOGRAPHIC, AND PATHOLOGICAL STUDY OF PULMONARY EMBOLISM BY J. G. MACLEOD AND I. W. B. GRANT From the Departments

Thorax (1954), 9 71.

A CLINICAL, RADIOGRAPHIC, AND PATHOLOGICALSTUDY OF PULMONARY EMBOLISM

BY

J. G. MACLEOD AND I. W. B. GRANTFrom the Departments of Medicine and Respiratory Diseases, University of Edinburgh

(RECEIVED FOR PUBLICATION MARCH 30, 1953)

This study of 60 cases of pulmonary embolismwas undertaken in an attempt to elucidate thepathological changes produced in the lungs and inthe pulmonary circulation in such cases and tocorrelate these with their clinical and radiographicmanifestations. The two aspects of the subjectwhich seemed to demand particular attention were

the diversity of the effects of pulmonary embolismand the difficulty in reconciling the variable andoften transient clinical and radiographic featuresof pulmonary infarction with the usual patho-logical conception of that condition, haemorrhagicpulmonary necrosis. This study is concernedprimarily with the effects of pulmonary embolism,and the events preceding embolism will be dis-cussed only in so far as they relate to subsequentdevelopments. The results of the study are

presented in two parts. First an account is givenof the clinical and radiographic findings and there-after the pathological interpretation is presentedin the light of these findings and with reference torelevant experimental and other work.

CLINICAL AND RADIOGRAPHIC FINDINGS

In the course of a period of approximately twoyears over 100 cases of pulmonary embolism were

seen in the medical, surgical, gynaecological andobstetric wards of the Western and EasternGeneral Hospitals, Edinburgh. Of these, only 60were considered suitable for detailed study. Inthe others, either the quality of the radiographswas poor, many of the films having perforce to betaken with a mobile apparatus, or radiographicexamination had not been made sufficiently oftenfor adequate observation of the course of thedisease.Of the 60 cases studied, two were suffering from

the effects of massive pulmonary embolism whenfirst seen. These were the only cases of this typeincluded in the series as all the others died beforeradiographic examination could be undertaken.The two cases were of considerable interest

radiographically and pathologically, and willbe considered separately. The remaining 58presented the features of pulmonary infarction.A diagnosis of pulmonary infarction can be

made with certainty only at necropsy. Duringlife it must depend on circumstantial evidence.It would have been possible to avoid presumptivediagnosis by confining the study to fatal cases.Additional material could then have been obtainedby studying in retrospect cases in which the con-dition was first recognized post mortem. Thispolicy was not adopted for two reasons. In thefirst place it was believed that clinical and radio-graphic observation of the non-fatal cases mightprovide information relevant to the pathology ofpulmonary infarction that could not be obtainedfrom the fatal cases alone. Secondly, a retrospec-tive clinical or radiographic study of cases notrecognized during life would almost certainly havebeen incomplete or unreliable. Accordingly it wasdecided to observe both fatal and non-fatal casesin which the diagnosis of pulmonary infarctionwas made during life, while admitting the possi-bility of inaccurate diagnosis in a few of the non-fatal cases.

In pulmonary infarction there is no singlemanifestation, clinical or radiographic, thatcannot occur in most other types of pulmonarydisease. In practice, however, a diagnosis ofpulmonary infarction can usually be made whenpleuritic pain, haemoptysis, or both, occur eitherin association with a known source for pulmonaryemboli, or in circumstances known to predisposeto thrombo-embolism such as prolonged recum-bency, a recent major surgical operation, or arecent confinement. Further support is given tothe diagnosis if the onset of symptoms suggestingpulmonary infarction is immediately preceded byan episode of acute dyspnoea, central chest pain,or circulatory collapse, such as might be producedby a fairly large embolus lodging in the pulmonarytrunk or in one of its major branches.

on June 11, 2020 by guest. Protected by copyright.

http://thorax.bmj.com

/T

horax: first published as 10.1136/thx.9.1.71 on 1 March 1954. D

ownloaded from

http://thorax.bmj.com/

J. G. MACLEOD and 1. W. B. GRANT

As one of the purposes of the investigation was

to elucidate the nature of the radiographicchanges in pulmonary infarction, the results ofthis examination were not allowed to influence theclinical diagnosis, except to exclude other diseases.It would, in fact, have been difficult to adopt any

other course as, at the time the investigation was

begun, we were unable, either in the light ofpersonal experience or from a study of the liter-ature, to formulate any reliable criteria for a

radiographic diagnosis of pulmonary infarction.

ANALYSIS OF CASES OF PULMONARY INFARCTIONThe number of cases in the series is too small

for detailed statistical analysis. Such an analysisin respect of predisposing circumstances and ofsymptoms such as pain and haemoptysis wouldalso be invalidated by the fact that these featureshave been employed as diagnostic criteria. Theanalysis which follows is thus intended to convey

mainly a general impression of the aetiologicalfactors, the clinical and radiological features, thecourse, the prognosis, and the complications in a

group of cases of pulmonary infarction.

AETIOLOGICAL FACTORS IN PULMONARY INFARCTIONSEX INCIDENCE.-The 58 cases consisted of 29

males and 29 females. This proportion supportsthe view of most observers that the disease isequally common in the two sexes.

CIRCUMSTANCES OF ONSET.-Twenty-five cases

occurred after a surgical or gynaecological oper-

ation and four after parturition. Twenty-ninecases, belonging to neither of these categories, are

referred to as "medical cases." Most of thesemedical cases developed pulmonary infarctionwhile under treatment in hospital for some otherdisease, but in four cases this was not so, pul-monary infarction itself being the reason foradmission to hospital. In each of these four cases

the infarction was preceded by venous thrombosisin a lower limb occurring when the patient was

ambulant and apparently in good health.RELATIONSHIP TO HEART DISEASE.-Of the 58

cases of pulmonary infarction, 17 had organicheart disease and 41 had not (Table I). Of the17, rheumatic heart disease was present in nine,and hypertension or coronary disease in theremainder. Heart disease was encountered in justover half of the " medical " cases, but in only twoof the 29 post-operative and post-partum cases.

AGE INCIDENCE.-Approximately two-thirds ofthe patients were between 40 and 70 years old.The average age, 55 years, was the same in both

post-operative and "medical" cases and also inall cases with, and in all cases without, heartdisease.

TABLE IRELATIONSHIP TO HEART DISEASE

Category of Case With WithoutCategryofCase Heart Disease Heart Disease

Post-operative 2 23Post-partum 0 4"Medical" 15 14

Total 17 41

SOURCE OF EMBOLUS.-As in most other reportson the subject, the commonest source was a lowerlimb vein (60% of all cases in this series). Therewas a higher proportion of cases in the heartdisease group in which the source of the emboluswas not identified. It may well be that in some ofthese cases the pulmonary infarction was due, notto embolism, but to primary pulmonary arterialthrombosis. Alternatively, the infarction mayhave been the result of embolism secondary tointra-cardiac thrombosis, for example, fromauricular fibrillation which was present in six ofthe 11 cases in which the source of the emboluswas not identified.

CLINICAL FEATURES OF PULMONARY INFARCTIONIn 52 of the 58 cases only one incident of pul-

monary infarction was observed. In five casesthere were two separate incidents weeks or monthsapart and in one case there were three such inci-dents. The survey of clinical features whichfollows is based on the total of 65 incidents. Asno important differences were noted in the clinicalfeatures of the post-operative, the post-partum,and the "medical " cases, an analysis has beenmade only in respect of the presence or absenceof heart disease.

PLEURITIC PAIN.-This was present in 58 of the65 incidents, giving a total incidence of 90%. Itwas usually the presenting symptom. The painwas on the right side in 33 cases, on the left in 16cases, and on both sides in nine cases. In sevencases there was no pain. There was no demon-strable difference in the incidence or distributionof pain between the cardiac and non-cardiacgroups.HAEMoPTYSIs.-Haemoptysis was present in 32

incidents (just under half) and absent in theremainder. It was seldom severe and often tookthe form of bloodstained sputum being expec-torated over a period of several days. Haemop-tysis was more common in patients withpulmonary congestion secondary to heart disease.

72



/T


ownloaded from


STUDY OF PULMONARY EMBOLISM

COUGH AND SPUTUM.-These (apart fromsputum associated with haemoptysis) were absentin more than two-thirds of all incidents.Frequently, when these symptoms were present,they antedated the pulmonary infarction andprobably had some other cause. The incidenceof cough and sputum was the same in the cardiacand non-cardiac groups. We came to the con-

clusion that the absence of cough and sputum lentsome support to a diagnosis of pulmonary infarc-tion.OTHER CLINICAL FEATURES.-Certain other

clinical features were seen.

Dyspnoea was a variable symptom and difficultto assess as it was so often associated withpleuritic pain. It seemed to bear little relation-ship to the size and number of the infarcts.

Pyrexia was variable from case to case. Mosthad a low fever of between 990 and 100° F. forthree to seven days, but in some there was no

pyrexia and in others the temperature rose, usuallyin the first 24 to 48 hours, to 102° or higher. Ina few cases there was low pyrexia persisting forone to four weeks. There was no recognizablecorrelation between the degree of pyrexia and theextent of the infarction.

Tachycardia.-The heart rate was above 100per minute in the majority of incidents. Usuallyit was abnormally rapid in proportion to thedegree of pyrexia, and in two of the seven inci-dents in which there was no pyrexia it exceeded120. In general, however, it helped little indiagnosis.

Cyanosis was present only in cases with veryextensive infarction and in cases with heartfailure.

Icterus, which Short (1952) observed in many

cases, was not noted in this series except inpatients with long-standing congestive cardiacfailure, in whom it was ascribed to cardiaccirrhosis.

Physical Signs in the Chest.-These wereseldom gross. The commonest abnormality was

the finding of coarse or medium crepitations,generally in association with diminished air entry.In a few cases the physical signs of consolidationwere present. In several patients physical signssuggesting pleural effusion were elicited, but inmany of these instances a subsequent radiographshowed that the signs were due, not to effusion,but to a high hemidiaphragm. In some cases thepleural rub and the radiographic changes were on

opposite sides. This observation was considered

of importance as it appeared to indicate that insome cases an infarct could be recognizable clinic-ally, by the pleural rub, and yet produce noradiographic changes.

Chaniges in Blood Count.-In most cases ofpulmonary infarction examination of the bloodshowed a polymorph leucocytosis. A commonfigure was 12,000 per c.mm., but total white countsof up to 17,000 per c.mm. were recorded even inthe absence of evidence of secondary infection.Significant degrees of anaemia were not observed.

Character of Pleural Liquid.-Pleural liquid wasobtained for examination in only four cases, as itdid not seem justifiable to submit patients withsmall effusions to the discomfort of needling. Inall four cases the liquid was serous in character,and the predominant cells were polymorphs andthe protein content was above 4%.

Cardiac Failure.-Congestive cardiac failurequite frequently preceded the development ofpulmonary infarction. In other cases the infarc-tion appeared to precipitate congestive cardiacfailure. Both these events were observed only incases with pre-existing heart disease. Acute corpulmonale developed in two cases of pulmonaryinfarction in which massive embolism occurred asa terminal event. No cases of chronic cor pul-monale resulting from multiple infarcts wereobserved.

RADIOGRAPHIC FEATURES OF PULMONARYINFARCTION

INCIDENCE AND DISTRIBUTION OF RADIOGRAPHICABNORMALITIES.-The first part of Table II showsthe distribution of the radiographic changes inrespect of the side or sides involved. Right-sidedinvolvement alone was commoner than left-sidedinvolvement alone, but bilateral involvement was

TABLE IIINCIDENCE AND DISTRIBUTION OF RADIOGRAPHIC

CHANGES

A. Radiographic Changes.Right side only.. 21lLeft side only 10 62 (95%)Bilateral 31JNone .. 3

65

B. Distribution of 117 Opacities in 65 IncidentsRight Side .. 68Upper zone .. .. 6Middle zone 12Lower zone 50

Left side .. 49Upper zone 4Middle zone .. .. 8Lower zone 37

73



/T


ownloaded from



the commonest of all (31 of 65 incidents). Inthree cases no changes were observed.

The second part of the table shows the dis-tribution of 117 opacities in the 65 incidentsaccording to the zone involved. The term " zone"is used because lateral films were not available inall cases. Some of these opacities were made up

of two or even three separate elements, which willbe analysed in Table III, but here only the zonaldistribution of the opacities is considered. Ifmore than one type of opacity was present in one

zone, only one was counted.The right lower zone, as in most other recorded

series, was the commonest situation for pulmonarvinfarcts and the left lower zone came next in orderof frequency. Infarcts were much less common

in the upper and mid zones.

NATURE OF THE RADIOGRAPHIC ABNORMALITIESIN PULMONARY INFARCTION.-It is possible toclassify most of the opacities seen as either ofpulmonary or pleural origin. There is little diffi-culty in recognizing pleural opacities due toeffusion or thickening from their characteristicappearance and distribution. Such opacitieswere present in many cases of this series. Atypical example is shown in Fig. 1.Homogeneous or patchy opacities of non-

pleural, presumably intrapulmonary, distributionconstitute the second common type of radio-graphic abnormality. Although, occasionally,opacities of this type correspond in position andshape with atelectatic segments the vast majoritydo not and are assumed to be produced by thepulmonary infarcts themselves. This inter-pretation has been confirmed in several cases ofthe present series at necropsy. For the purposesof description this type of opacity will be referredto as a "pulmonary opacity." It resembles thatfound in pneumonia and may have either a seg-mental or a multilobular distribution. Fig. 2shows bilateral pulmonary opacities due to infarc-tion.The third type of abnormality, which will be

referred to as a "linear opacity," is perhaps themost characteristic of all (Figs. 1, 3, and 4). Thisterm has been applied to horizontal or obliquelines which at first are thick, but which laterbecome thin and sharply defined. In many of thepresent cases the linear opacity was noted within24 hours of the first symptoms, but usually it didnot appear until after the third or fourth day.Usually it disappeared within two to three weeks,sometimes in less than one week, but occasionallyit persisted for several months.

The sharply defined triangular opacity whichwas once said to be a common feature of pul-monary infarction was not observed in any casein this series.Another radiological abnormality commonly

found in these cases of pulmonary infarction wasan alteration in the position and function of thediaphragm (Figs. 3 and 4). The commonestfeature in the acute stage was a difference in levelbetween the two domes, with inhibited movementof the higher dome. This abnormality is oftenobserved in pleurisy from other causes, but wehave gained the impression, which remains to beconfirmed by further investigation, that theelevation of the hemidiaphragm with pulmonaryinfarction is greater and of longer duration thanwith pleurisy from other causes. True paralysisof one hemidiaphragm, with paradoxical move-ment on sniffing, was noted in four of the 58cases. Although this can occur with pleurisy fromother causes the incidence must be very muchlower than in this series of cases of pulmonaryinfarction. In none of the four cases has dia-phragmatic function so far returned. Diaphrag-matic adhesions were observed frequently afterrecovery from pulmonary infarction.The radiographic features observed in this

series of cases of pulmonary infarction may besummarized as follows: (a) Pleural opacities,(b) pulmonary opacities resembling those foundwith pneumonic or bronchopneumonic con-solidation, (c) linear opacities, (d) elevation of ahemidiaphragm with inhibited movement or,occasionally, true paralysis, (e) the frequentlybilateral distribution of lesions.

It must be emphasized that the individual typesof radiographic abnormality encountered inpulmonary infarction are themselves non-specific.Each one of them can occur, for example, incases of pneumonia with pleurisy. The claim,however, is made that the coexistence of two ormore of these abnormalities, particularly if thechanges are bilateral, constitutes a radiologicalpicture which gives strong support to a clinicaldiagnosis of pulmonary infarction. The pictureoften alters rapidly either as a result of changesin existing opacities, particularly in linear shadows,or from the appearance of new opacities due tofurther infarction. Radiographs repeated atintervals of four or five days after pulmonaryembolism are therefore helpful in diagnosis.

INCIDENCE OF VARIOUS TYPES OF RADIOGRAPHICABNORMALITY.-Table III shows the frequency ofeach of the three main types of opacity. Onehundred and forty-nine separate opacities were

74



/T


ownloaded from


STUDY OF PULMONARY EMBOLISSA7

. . . ~~~~~~~~~~~~~~~~~~~~~~~......ll.~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~.t.....O1.-Bilateral pleural opacities; pulmonary opacity right lower

zone and linear opacities left lowei zone.

FIG. 3.-High right diaphragm; bilateral linear opacities.

observed in the course of 65 incidents. Theirrelative frequency differed in cases with heartdisease as compared with cases without heartdisease. In the cardiac group pulmonary andpleural opacities were frequently encountered.Possibly in some cases the pleural opacity was duenot to the effects of pulmonary infarction but to

s- ...._! ................................ RM.

.. M _' .. ' ..s. tb: _

i ::. tw ....: :t :. .>_.._ - a: : .. :g gt.. _w ..

.i_ ..,,:_; _rF .n:. . b. s-__. ::q._-is_- ...._B _ Be.:._vv _b,s . q_R. ....._r 't .'.1.. .'.:_:.-_aF ......._R8: ., o.

L :.>*_E.: Y':.. ::. .... .. ::::.

&.i:;° X q i.

£:

t ...

g .§.,.:0. ,.:. X......

TFIG. 2.-Bilateral pulmonary opacities resembling consolidation.

FIG. 4.-High diaphragm and linear opacity on right side.

a small pleural transudate secondary to cardiacfailure. Linear opacities were uncommon in thisgroup of cases.

In the non-cardiac group the position wasdifferent in that linear opacities were foundas frequently as pleural opacities and morefrequently than pulmonary opacities. It is not

Wil

75



/T


ownloaded from


J. G. MACLEOD and I. W. B. GRANT

clear why this should be so, but a possibleexplanation is that the pathological nature ofwhat is regarded clinically as pulmonary infarctionis in some way different in the cardiac and non-cardiac groups.A hemidiaphragm was definitely elevated in

almost 50% of all incidents of pulmonary infarc-tion. It was not possible to draw any reliableconclusion regarding the relative frequency of thisabnormality in the cardiac and non-cardiacgroups.The figures in the present series may be con-

trasted with those of Short (1951), who claimedto demonstrate an " infarct shadow" in 88 %,pleurisy in 56%, and elevation of the diaphragmin 39%. Short included what we have called"pulmonary opacities" and "linear opacities"under the heading of infarct shadows.

TABLE IIINATURE OF 149 SEPARATE OPACITIES IN 65

INCIDENTS

Cases with Cases withoutAll Cases Heart Heart

Disease Disease

Pulmonary opacity 47 (31%) 14 (41%) 33 (28%)Pleural opacity 59 (40%) 18 (53%) 41 (36%)Linear opacity 43 (29%) 2 (6%) 41 (36%)

RATE OF CLEARING OF RADIOGRAPHIC OPACITIES.-The average time taken for resolution in caseswith heart disease (48 days) was more than doublethat in cases without heart disease (23 days). Fatalcases and cases in which a complete series ofradiographs was not obtained have been excludedfrom this analysis.

In many cases of the non-cardiac group radio-graphic resolution was extremely rapid, completeclearing of the opacity or opacities taking placein less than a week.

CLINICAL AND RADIOGRAPHIC FEATURES OFCARDIAC AND NON-CARDIAC GRouPs.-Therewere three clear differences between thosetwo groups of cases of pulmonary infarc-tion. (1) Haemoptysis was commoner in thecardiac group if pulmonary congestion waspresent. (2) Linear opacities were much com-moner in the non-cardiac group. (3) Clearing ofthe radiographic opacities was more rapid in thenon-cardiac group.

DIFFERENTIAL DIAGNOSIS OF PULMONARYINFARCTION

Pulmonary infarction may be mistaken formany other diseases. In this series the initialdiagnosis included pulmonary tuberculosis,

primary pleural effusion, pneumonias of varioustypes, lung abscess, and bronchial carcinoma. Itmay be difficult to distinguish pulmonary infarc-tion from post-operative atelectasis. In the lattercondition phlegm is present though it may not beproduced, pleuritic pain is not prominent, and theradiographic opacities, which have a segmentaldistribution, usually clear promptly after posturalcoughing or bronchoscopic aspiration.

PROGNOSIS OF PULMONARY INFARCTIONThe total mortality in the series was 21%.

There was no striking difference in mortality be-tween the post-operative and " medical" groups.There was a higher mortality in cases with heartdisease than in cases without, but this differencewas almost certainly related to the heart diseaseper se.Acute cor pulmonale occurred in two cases in

which the small embolism causing infarction waslater followed by a massive embolism. Both thesepatients died.

Five cases developed pulmonary suppuration,presumably as a result of secondary bronchogenicinfection of infarcted lung, and of these casesfour died. Three of the five patients, the one whosurvived and two of those who died, developedempyema as a further complication. These casesillustrate the importance of effective chemotherapyas a prophylactic measure in the treatment of pul-monary infarction.While many cases had small localized pleural

opacities, in only three did a large pleuraleffusion develop. In these cases the effusion tooksome months to absorb and there was considerableresidual pleural thickening, but the patients allsurvived. Pulmonary infarction should be keptin mind as an occasional cause of a large chronicpleural effusion.

Six cases, all with pre-existing heart disease,developed congestive cardiac failure followingpulmonary infarction, and of these six cases fourdied.

MASSIVE PULMONARY EMBOLISMTwo fatal cases of massive pulmonary embolism

were studied clinically and pathologically. Onepatient died 10 hours after the embolism, whilethe other survived for six weeks after the onset ofsymptoms. They each presented a number ofinstructive features relating to the effects of pul-monary embolism.

CASE 1.-A woman, aged 59 years, suffering fromthrombo-phlebitis, developed a massive pulmonaryembolism. A radiograph of the chest was obtained nine

76



/T


ownloaded from



hours after the initial symptoms and one hour beforedeath. It showed that the vascular markings of the rightlung were distinctly less prominent than those of theleft, and that the markings on the left were less prominentthan on a routine radiograph taken six weeks previously.In addition, there was marked distension of the superiorvena cava. At necropsy a coiled embolus totally occludedthe lumen of the right pulmonary artery and partiallyobstructed the left pulmonary artery. The lungs them-selves appeared normal both macroscopically andmicroscopically.

This case illustrates the radiographic pheno-menon first described by Westermark (1938),namely, relative ischaemia of a lung or part of a

lung distal to an embolism. He also included inhis description a dense pulmonary artery shadowwith a sharp distal termination, presumablyrepresenting the embolus in situ.CASE 2.-A man, aged 36, was admitted to hospital

two months after internal saphenous ligation for varicoseveins, with a two weeks' history of increasing breathless-ness on exertion which had started suddenly after a shortperiod of central chest pain.On admission he was orthopnoeic, deeply cyanosed,

and mildly febrile. The heart rate was 84 per minute andthe rhythm regular; there was no clinical evidence ofcardiac enlargement and no peripheral oedema, butgallop rhythm was present and the external jugular veinswere distended. Radiography of the chest five daysbefore admission showed prominence of both hilarshadows. The lower branch of the right pulmonaryartery terminated abruptly at the lower end of the hilumwithout branching in the normal manner, and only afew spidery vascular markings were visible in the lowerzone of the right lung. The markings in the left lowerzone were also less prominent than normal. The reduc-tion in the vascular markings in the lower lung fieldsresembled that found in emphysema. An electrocardio-gram showed an abnormal pattern of the type associatedwith myocardial ischaemia.

Despite treatment with anticoagulants and digitalis,the breathlessness increased and gross peripheral oedemadeveloped. A radiograph three weeks after admissionshowed marked cardiac enlargement, mainly rightsided with prominence of the pulmonary trunk, inaddition to the abnormalities noted in the previous film.Five days later he had a severe attack of upper chestpain, situated mainly to the left of the mid-line and notrelated to respiration, which was followed by a furtherincrease in dyspnoea and cyanosis. He died on thefollowing day.At no time during the illness were crepitations heard

on auscultation of the lungs. A further electrocardio-gram, taken a few hours before death, showed changescompatible with acute right ventricular strain.At necropsy the right and left pulmonary arteries and

their larger branches were all completely occluded byfirm, adherent thrombus.The lungs were emphysematous and there was an early

infarct in the right lower lobe. There was gross dilatation

and some hypertrophy of the right ventricle and of theright atrium. Extensive venous thrombosis was found inthe lower limbs.

This patient presented the clinical, radiographic,electrocardiographic, and pathological features ofsubacute cor pulmonale which was considered to be theresult of progressive thrombosis throughout thepulmonary arterial tree initiated by a massive embolism.Despite the severity of the vascular occlusion, pulmonaryinfarction was not a conspicuous feature. Some of theradiographic features of pulmonary embolism describedby Westermark were present, but these could equallywell have been produced in this case by emphysema.

AN INTERPRETATION OF THEPATHOLOGICAL AND OTHER PROCESSESFOLLOWING PULMONARY EMBOLISM

In considering the effects of pulmonary embo-lism attention must first be paid to the size of theembolus. Gibbon, Hopkinson, and Churchill(1932) have demonstrated in experimentalanimals that 60% of the total pulmonary circu-lation must be obstructed before the bloodpressure falls and the right ventricle is embar-rassed, and probably 80% before death is likely.It is after such massive emboli that acute cor pul-monale is most frequently encountered. Thepresent series included four cases of this type. Intwo of them the fatal embolism was preceded bypulmonary infarction due to smaller emboli.A less massive embolus may lodge near the

bifurcation of the pulmonary trunk and causeacute dyspnoea, circulatory collapse, or centralpain simulating myocardial infarction and prob-ably due to myocardial ischaemia. These symp-toms may be relieved if the embolus fragmentsand passes on to the right and left pulmonaryarteries. This may then result in bilateral infarc-tion with pleuritic pain, as occurred in one of ourcases. Smaller emboli usually produce pulmonaryinfarction without any preceding acute circulatorydisturbance.

In this as in other series the right lower lobeand the left lower lobe were in that order themost frequent sites of infarction. This distribu-tion of embolic lesions has been reproduced inexperimental animals (Businco and Cardia, 1931)by the injection of bismuth into a femoral vein.Subsequent fluoroscopic examination of the lungsshowed the opaque medium filling the right lowerlobe first and then the left lower lobe before beingdistributed to the rest of the lungs. This has beenexplained as being related to the direction, length,width, and distribution of the right pulmonaryartery.

In many of our cases the infarction wasbilateral. The probable explanation of this is that

77



/T


ownloaded from


J. G. MACLEOD and I. W. B. GRANT

the clot, which must be friable by the very reasonof its detachment, is liable to disintegrate on itsway to the lungs. The diagnostic importance ofthese bilateral lesions has already been empha-sized.

PULMONARY INFARCTIONEmbolism without infarction is well recognized

in experimental animals in which it is very diffi-cult to induce infarction except in the presence ofpulmonary congestion. In some cases of pul-monary embolism in man there may be no infarc-tion. This occurs if the embolus does not whollyocclude the vessel, as when a coiled emboluslodges in a large pulmonary artery, or if the col-lateral circulation is adequate to prevent infarc-tion. In most cases, however, embolism is fol-lowed by infarction. The lesions produced varygreatly in extent and severity from case to case.

HAEMORRHAGIC INFARCTIONIn this, the most severe type of pulmonary

infarction, the histological appearance is that ofhaemorrhagic necrosis (Figs. 5 and 6). In one areaof Fig. 6 the alveolar walls can just be detected,but even here it is clear that they have beendamaged beyond repair. In the present serieswhen infarcts of this type were demonstrated atnecropsy, generalized pulmonary congestion waspresent in most instances.

INCOMPLETE INFARCTIONIn haemorrhagic infarction the alveoli are

destroyed and repair takes place by organization.Such a slow fibrotic process is not compatible withthe variable and transitory radiographic findingsin the majority of cases of this series. There wasrarely any residual scar demonstrable such aswould be expected if a haemorrhagic infarct hadoccurred and healed by fibrosis. Although experi-mentally it is difficult to produce haemorrhagicinfarction in normal lungs, clinical and radio-graphic studies such as this show that some changedoes occur in the healthy human lung as a resultof embolism. Pathological material to demon-strate this process is difficult to obtain as thesepatients generally survive. Hampton and Castle-man (1940) drew attention to certain lesions whichoccurred as a result of embolism in experimentalwork carried out in animals with healthy lungs.These were not regarded as true infarcts becausethere was no destruction of the alveolar wall.There was, however, local pulmonary congestionand oedema with some haemorrhage into thealveoli. Identical findings were also demonstrated

in a woman who had a post-operative infarct andwho died " a few days later " from massive pul-monary embolism. To these changes Hamptonand Castleman applied the term " incompleteinfarction."

In the pathological material available in thepresent series examples of incomplete infarctionwere readily found. Figs. 7 and 8 show what isconsidered to be the earliest stage of incompleteinfarction, i.e., localized pulmonary congestion.It is seen that this is most marked in the pleuraand subjacent lung. In contrast with truehaemorrhagic or "complete" infarction seen inFig. 5 the alveolar walls are intact.A lung section from the patient (Case 2) who

had bilateral pulmonary emboli but no macro-scopic evidence of infarction shows congestionand considerable oedema (Fig. 9), which can beregarded as a slightly later stage of incompleteinfarction. A similar example is shown in Fig. 10.This patient had widespread pulmonary infarctiondemonstrated radiographically; shortly beforedeath this was clearing dramatically following therelief of generalized pulmonary oedema by mer-curial diuretics. At necropsy only one infarctedarea was found; this showed localized congestionand oedema and was presumably capable ofresolution.A more advanced stage of incomplete infarction,

where in addition to congestion and oedema thereis alveolar haemorrhage, is seen in Fig. 11. Thealveolar walls, however, have the appearance ofbeing viable. Haemoptysis is to be expected ina lesion of this type but not in the previousexamples of incomplete infarction. Fig. 12 showsa late stage of incomplete infarction, the changesof which may or may not be reversible. At thisstage incomplete infarction merges into the com-plete. The pathological changes in pulmonaryinfarction thus range from congestion to oedema,alveolar haemorrhage, alveolar wall destruction,and complete haemorrhagic necrosis. In almostevery case of complete infarction in this seriesstudy of adjacent areas of lung showed all degreesof incomplete infarction.

PLEURAL CHANGES IN PULMONARY INFARCTION

Clinical, radiographic, and pathological evidenceall suggest that in infarction the pleura andadjacent lung are predominantly involved. Pleur-itic pain is the commonest symptom. In thesurvey of radiographic findings it was concludedthat the changes most frequently seen are pleuralin origin. Hampton and Castleman have shownthat the pleura is always involved, and by careful

78



/T


ownloaded from


STUDY OF PULMONARY EMBOLISM 79

FIG. 7.- x 3. Incomplete infarction showing congestion of pleuraland sub-pleural areas.

FIG. 5.- x 3. True haemorrhagic or complete infarction seen bestat the top of the section. The lowest dark area is an artefact.,

FIo. 6.- x 50. High power view of Fig. 5, showing extensive FIG. 8.- x 50. High power view of Fig. 7, showing congestion ofdestruction of the alveolar walls. sub-pleural areas.

F



/T


ownloaded from


J. G. MACLEOD and 1. W. B. GRANI

.t

FIG. 9.- x 50. Incomplete infarction showing congestion and FIG. 1.-X 100. Incomplete infarction showing alveolar haemor-oedema. rhage.

Fici. 10.- x 50. Incomplete infarction showing congestion andoedemna.

FIG. 12.- x 50. Late stage of incomplete infarction showing alveolarwall destruction.

80



/T


ownloaded from


STUDY OF PULAIONARY EMBOLISM

dissection of infarcts have demonstrated that thelongest axis of the infarct is parallel to the longestpleural surface involved. In the earliest stagesof incomplete infarction the congestion is mostmarked in the sub-pleural areas (Fig. 7). A pos-sible explanation of this is that the degree ofischaemia is more severe at the periphery of thelung, where the collateral circulation is likely tobe less efficient.

CIRCULATORY ADJUSTMENTS IN COMPLETE ANDINCOMPLETE INFARCTION

Until recently the experimental evidence regard-ing the relative importance of the bronchial andpulmonary circulation in the production of infarc-tion has been conflicting. Ellis, Grindlay, andEdwards (1951, 1952) have done much to clarifythe position by experiments in dogs using new

techniques which appear reliable. They demon-strated that the bronchial artery to a lobe couldbe occluded without resulting pathological changesin that lobe, and concluded that the pulmonaryarterial circulation supplied the nutritive require-ments of the lung beyond the hilum. Occlusionof a branch of the pulmonary artery as well as thebronchial artery produced no pathological change,i.e., an adequate collateral circulation was pro-vided by pulmonary capillaries. They showed,however, that when the bronchial circulation wasintact emboli resulted in complete and incompleteinfarcts, and that the amount of haemorrhagevaried with the degree of dilatation of thebronchial vessels. Other workers have shownthat there is an enormous dilatation of the bron-chial arteries beginning a few days after infarc-tion. Ellis therefore makes two main claims:(1) The pulmonary arteries provide the collateralblood supply after infarction, and (2) the haemor-rhage in infarction comes from the bronchialarteries. With this work in mind we suggest thefollowing scheme (see next column) as compatiblewith the pathological changes.While it is believed that the interpretation given

applies to embolism throughout the greater partof the pulmonary tree, it may not apply to embo-lism involving a major pulmonary artery. Herebroncho-pulmonary arterial anastomoses mayalter the circulatory adjustments (Ellis and others,1952).Complete infarction is particularly liable to

occur when there are adverse circulatory factorssuch as generalized pulmonary congestion. Thisconcurs with the repeated observation that inanimals pulmonary congestion is a prerequisite fortrue infarction. In the present series it has been

OCCLUSION OF PULMONARY ARTERY BY EMBOLUS

Pulmonary ischaemiaI ~~I

Collateral pulmonary Collateral pulmonary Collateralblood supply inadequate blood supply not quite pulmonary blood(e.g., in presence of adequate supply adequategeneralized pulmonarycongestion)

Prolonged ischaemia Anoxia but no No alveolarresulting in destruction destruction of damageof alveolar walls alveolar walls

Influx of blood from Influx of blood from No significantdilating bronchial dilating bronchial bronchial arteryarteries throughout arteries largely confined dilatationdamaged area to vessels but produ-

cing congestion,oedema, and some

alveolar haemorrhageCOMPLETE INCOMPLETE EMBOLISMINFARCT INFARCT WITHOUT

INFARCTION+I

Healing by fibrosis Healing by absorption

demonstrated that infarction in patients with heartdisease runs a longer course compared with infarc-tion in healthy lungs; this is presumably becausemost of these cases had complete infarcts.

It is suggested that incomplete infarctionresulting in local congestion and oedema of lungand pleura gives a satisfactory explanation of theoutstanding clinical and radiographic features ofembolism in healthy lungs. Complete infarction,though it may occur in healthy lungs, is muchmore common if the pulmonary circulation is in-efficient. If this is true it is obvious that, in thetreatment of pulmonary infarction, particularattention must be paid to the relief of anygeneralized pulmonary congestion.

CORRELATION OF CLINICAL AND PATHOLOGICALFINDINGS

PULMONARY OPACITIES.-In several cases of thepresent series opacities of this type were shown atnecropsy to have been produced by the infarctitself. It has been clearly demonstrated by manyprevious workers that the infarct shadow is notnecessarily triangular in shape. This point, how-ever, is worthy of further emphasis. The infarctshadow is often ill defined, as indeed would beexpected if it reflects local congestion and oedemaalone or bordering on an area of complete infarc-tion. We have often been impressed by thecontrast between the radiographic evidence ofextensive pulmonary involvement and the smallmacroscopic infarct found at necropsy. Theexplanation suggested is that around the areaof destruction there is extensive congestion andoedema, i.e., the changes of incomplete infarctionsurround the complete infarct. In cases with

81



/T


ownloaded from



heart disease the infarct shadow is more clearlydefined as would be expected when completeinfarction predominates.

PLEURAL OPACITIES.-In most cases thesemust represent small pleural effusions or areas offibrinous exudate overlying a pulmonary infarct.

LINEAR OPACITIES.-It is suggested that theseare produced by sharply localized lesions of thepleura and subjacent lung. This explanation isin keeping with the observation that linear

shadows are more commonly associated withpleural than with pulmonary opacities. Of a totalof 43 linear opacities in the present series, 37 wereaccompanied by pleural opacities on the same,

opposite or both sides; in contrast only six of 43linear opacities were associated with pulmonaryopacities. As there was little difference betweenthe total numbers of pleural and pulmonaryopacities in the series these figures suggest thatboth the linear and pleural opacities may well bemanifestations of the same process, viz., pleuralcongestion, oedema and exudation, and that thelinear opacity represents a localized pleural re-

action over the site of a pulmonary infarct whichmay itself not be visible. A further point infavour of a pleural origin for linear opacities isthat in several cases a pleural opacity was observedto change into a typical linear opacity.

It has not been possible to obtain with certaintya histological section of a linear opacity. This isprobably because the linear opacity is usually a

transitory phenomenon and occurs most frequentlyin non-fatal cases. Occasionally, however, a

linear shadow does not resolve. Hampton andCastleman were able in one such instance tocorrelate the clinical, radiographic, and necropsyfindings. They demonstrated convincingly thatthis linear opacity was due to pleural thickeningwhich had proceeded to fibrosis. Hampton andCastleman were reluctant to accept a pleuralorigin for all linear opacities and suggested thealternative of discs of atelectasis. Short (1951)also supported this view. It is difficult to see whyatelectasis should occur when there is no patho-logical evidence of bronchial or bronchiolarobstruction and when the production of phlegmin infarction is usually inconspicuous.The above evidence of a pleural origin for linear

opacities is as yet incomplete, but the problem isbeing subjected to further study. If a pleuralorigin for linear shadows is accepted the relativeinfrequency of such shadows in patients withheart disease may be related to the completeinfarction which usually occurs in this type of

case, the pleural shadow being obscured by thepulmonary shadow.DIAPHRAGMATIC CHANGES.-(a) Immobility of

the diaphragm is probably reflex in origin andassociated with the pleuritic pain. (b) Paradoxicalmovement was seen in four cases, and was appar-ently permanent. Its cause is unknown. It isconceivable that the phrenic nerve trunk isinvolved in a pleural reaction. Possibly thebranches of the nerve as they enter the diaphragmare damaged by pleurisy. This may be oneexplanation of so-called "idiopathic'" diaphrag-matic eventration, i.e., an end-result of pulmonaryembolism. (c) Local impairment of diaphragmaticmovement occurs as a result of pleural adhesions.

. SUMMARYThe clinical and radiographic features of 60

cases of pulmonary embolism are presented; ofthese, 58 had pulmonary infarction and two hadmassive pulmonary emboli without infarction.

After pulmonary embolism, there may be noinfarction, incomplete infarction, or completeinfarction. Consideration is given to the partsplayed by the pulmonary and bronchial circu-lations in these events.

Incomplete infarction is characterized bylocalized congestion and oedema of the lungand pleura proceeding in some instances to intra-alveolar haemorrhage. This is the manifestationof embolism most frequently encountered inhealthy lungs.Complete or haemorrhagic infarction is charac-

terized by necrosis of the alveolar walls and is mostcommonly found in the presence of generalizedpulmonary congestion.The radiographic appearances are described as

pulmonary, pleural, and linear opacities, anddiaphragmatic abnormalities. The right lowerzone is most commonly involved. The distributionof the radiographic lesions is frequently bilateral.Linear opacities are frequently associated withpleural opacities and are much commoner inpatients without heart disease. They have ten-tatively been attributed to pleural and subjacentpulmonary oedema rather than to plaques ofatelectasis. The resolution of the radiographicchanges is more rapid in healthy lungs than in thepresence of generalized pulmonary congestion.This has been attributed to the higher incidenceof incomplete infarction in the former circum-stances, and of complete infarction in the latter.Radiographs repeated at intervals of four or fivedays are valuable in the diagnosis of pulmonaryinfarction.

82



/T


ownloaded from



Bronchogenic pulmonary suppuration is a com-mon and often fatal complication. Pulmonaryinfarction may cause a large, persistent pleuraleffusion or may precipitate cardiac failure. Per-manent paralysis of a hemidiaphragm may followpulmonary infarction.

In treatment prophylactic chemotherapy andmeasures directed to the relief of generalized pul-monary congestion are important.The diverse effects of pulmonary embolism are

explained by the variation in the size and numberof the emboli, by the frequently bilateral distribu-tion of the resulting infarcts, by the presence orthe absence of pulmonary congestion before theembolism, by the type of infarct that develops,

complete or incomplete, and by complicationssuch as infection.The authors are indebted to their colleagues in the

Northern Group of Hospitals, Edinburgh, who sowillingly and freely gave them access to patients, andto Professor L. S. P. Davidson, Professor John Crofton,and Dr. R. W. D. Turner for much helpful criticism.

REFERENCESBusinco, O., and Cardia, A. (1931). Fortschr. Rontgenstr., 44, 60.Ellis, F. H., Jr., Grindlay, J. H., and Edwards, J. E. (1951). Surgery

30, 810.

--- 5(1952). Ibid., 31, 167.Gibbon, J. H., Hopkinson, Mary, and Churchill, E. D. (1932). J.

clin. Invest., 11, 543.Hampton, A. O., and Castleman, B. (1940). Amer. J. Roentgenol.,

43, 305.

Short, D. S. (1951). Quart. J. Med., 20, 233.- (1952). Brit. med. J., 1, 790.Westermark, N. (1938). Acta radiol., Stockh., 19, 357.

83



/T


ownloaded from


Documents

CLINICAL, RADIOGRAPHIC, AND PATHOLOGICAL STUDY OF ... · A CLINICAL, RADIOGRAPHIC, AND PATHOLOGICAL STUDY OF PULMONARY EMBOLISM BY J. G. MACLEOD AND I. W. B. GRANT From the Departments