Postgrad Med J: first published as 10.1136/pgmj.38.435.13 on 1 January 1962. Downloaded from

POSTGRAD MED. J. (I962), 38, 13

PULMONARY EMBOLISM A-ND THROMBOSIS R. MARSHALL, M.D., M.R.C.P. Nuffield Department of Surgery, Radcliffe Infirmary, Oxford

SUDDEN death following a call for a bedpan is a chest pain and ECG changes when the balloon classical mode of presentation of pulmonary em- was inflated in a smaller artery (Brofman, Charms, bolism, but it is probable that cases of this severity Kohn, Elder, Newman and Rizika, I957). On the form only a small proportion of the total. Pul- other hand, it is a common practice to impact the monary embolism may be caused by blood clot, tip of a cardiac catheter in branches of the pul- fat, cells or foreign bodies, but most commonly by monary artery of about 2 mm. diameter with no blood clot. ill effects. In spite of the undoubted ability of emboli to produce reflex effects under certain cir- cumstances, there is a growing body of evidence The Physiological Effects that the effects of pulmonary embolism by inert The physiological effects ofpulmonary embolism particles are produced by the magnitude of the depend upon the size and number of the emboli mechanical blockage of the pulmonary vessels and and they may produce these effects either by are not dependent upon the size of the emboli or mechanical blockage of the circulation when they reflex vasoconstriction (Daley, I957; McEvoy, are of sufficient size or number, or by the initiation Harder and Dale, 1958; Williams, 1956). Comroe, of reflex broncho- or vasoconstriction. This point van Lingen, Stroud and Roncoroni (I953) have by copyright. is of considerable clinical importance because of suggested that some of the effects of pulmonary the possibility that reflex effects may be reduced embolization may be caused by serotonin liberated by drug treatment, whereas mechanical blockage is from the platelets in a thrombus. Serotonin is less likely to be affected by drugs. There has been one of the few substances with a strong vaso- much controversy over the part played by reflex constrictor action on the pulmonary arteries. action in causing the symptoms and disability of Following pulmonary embolization of dogs with pulmonary embolism. Removal of one lung in autologous blood-clot the level of serotonin in the man has no effect on the pulmonary arterial arterial blood has not been found to be raised pressure and causes very little respiratory dis- (Sanders, Waalkes, Gilbert and Terry, I959), but http://pmj.bmj.com/ ability. Dogs are able to live with less than 25% this is probably an insensitive way of detecting the of their pulmonary vascular bed remaining. These liberation of serotonin in the lungs. Work in observations have led people to suppose that in progress suggests that the cardiovascular effect of clinically severe pulmonary embolism the throm- embolization by platelet agglutinates is greater bus itself is unlikely to block so much of the than that of an equivalent mass of blood clot and pulmonary vascular tree that it can cause right this difference may be due to the serotonin heart failure. It has been supposed that reflexes, liberated. initiated by impaction of emboli in the smaller With this physiological evidence in mind it is on September 24, 2021 by guest. Protected arteries, cause reflex vasoconstriction of the whole useful to consider the clinical picture. Small of the pulmonary vessels. Much experimental emboli, if they do not produce reflex vasoconstric- work has been done to try to determine the role tion, will cause no symptoms and give rise to no played by reflex action in pulmonary embolism. signs; they will therefore pass unnoticed. A Reflex vasoconstriction has been demonstrated in single large embolus will have to obstruct more the contralateral lung of dogs when small glass than one main pulmonary artery if it is to cause beads are injected into the other lung, care being symptoms in a subject with normal Lings. This taken that none of the beads escaped from the raises the problem of the importance and the embolized lung (Niden and Aviado, I956). Most incidence of repeated small pulmona-ry e-nboli of the investigators attribute reflex activity to which could in time obstruct so much of the emboli which lodge in vessels of ioo t or less in pulmonary vascular tree that heart failure r sulted, diameter, i.e. in the arterioles. In man a balloon or could quietly reduce the size of the pulmonary inflated in the main pulmonary artery produces no vascular bed so that when a massive pulmonary reflex effects, but an instance has been reported of embolism did occur it could bie fatal. &l Postgrad Med J: first published as 10.1136/pgmj.38.435.13 on 1 January 1962. Downloaded from

14 POSTGRADUATE MEDICAL JOURNAL _7anuary I962 Detection of Emboli large embolus and the blood flow to an embolized Because of the clinical importance of silent, region may start to increase only a few minutes repeated pulmonary embolism it would be of value after the embolus impacts. to know when and under what circumstances silent Measurement of the diffusing capacity of the pulmonary embolism occurs and to what extent lungs may provide another method for the detec- the body's defence mechanisms are able to deal tion of pulmonary embolism. The diffusing with the emboli. Pulmonary emboli may not be capacity ofthe lungs is most conveniently measured detectable clinically or by X-ray or ECG examina- as the diffusing capacity for carbon monoxide and tion and other methods of investigation may have it is a measure both of the permeability of the to be used. A method which promised to be of alveolar membrane to carbon monoxide and of the some use was the measurement of the dif- surface area of the alveolar membrane in contact ference in carbon dioxide tension (Pco2) between with blood in the capillaries. If the blood supply the end tidal gas and the arterial blood. In a to some of the alveoli is cut off the diffusing subject with uniform distribution of gas in his capacity will be reduced and the reduction in dif- lungs the air sampled at the end of a normal fusing capacity, in the absence of other lung breath is representative of the gas from the alveoli disease, nm,ay give an indication of silent pulmonary and this should have the same Pco2 as the embolism. The single breath method of measure- arterial blood, since the alveolar wall presents no ment of the diffusing capacity for carbon monoxide diffusion barrier to carbon dioxide. If some of the is probably particularly useful for this purpose alveoli have no blood supply, because the arteries because the full inspiration which is necessary for have been blocked by emboli, then these alveoli this test probably minimizes any inequality of will contain no carbon dioxide and the gas from ventilation which might be caused by the emboliza- these alveoli will dilute the carbon dioxide in the tion. The usefulness of the method would depend gas coming from alveoli perfused with blood and on the detection of a fall in diffusing capacity on so the end tidal (alveolar) Pco2 will be less than repeated measurements; the normal range is too that measured simultaneously in the artery. wide for a single measurement to be of value. by copyright. This method, suggested by Severinghaus (Sever- inghaus and Stupfel, 1957), could, in theory, give Clinical Features a measure of the proportion of alveoli from which Small or medium-sized pulmonary emboli may the blood supply has been cut off. Experimenta be completely 'silent'. Larger emboli, which work on dogs (Julian, Travis, Robin and Crump, cause gross obstruction of the pulmonary circula- I960) and observations in man (Robin, Forkner, tion, cause the sudden onset of symptoms. Dys- Bromberg, Croteau and Travis, I960) have sug- pncea is often the main symptom and may be the gested that the method may indicate the extent of only one. A history of sudden attacks of dyspncea pulmonary artery obstruction following embolism. is often obtained from patients who have had More recently Severinghaus, Swenson, Finley, recurrent emboli. The cause of the dyspncea, http://pmj.bmj.com/ Lategola and Williams (i96i) have found that the as of the dyspnea secondary to pulmonary hyper- difference between end tidal and arterial Pco2 is tension, is not known. It may be reflex from dis- much less than would be expected from the tension of the right side of the heart or it may amount of pulmonary circulation obstructed. The result from inadequate blood supply to the respira- reason for this is a compensatory mechanism by tory muscles. Pain in the chest may be of two which abolition of the blood supply to a part of types. First, massive embolism may cause sub- the lung causes a reflex bronchoconstriction which sternal pain or discomfort similar to that of myo- on September 24, 2021 by guest. Protected reduces the ventilation to that part. Measurements cardial infarction. This may arise from distension of the end tidal, arterial Pco2 difference before of the pulmonary artery or may be of myocardial and after experimental pulmonary embolism of origin due to reduction in coronary blood flow. dogs with blood thrombi (Marshall, Allison, Bos- Substernal pain is not necessarily of cardiac origin; man, Sabiston and Dunnill, in preparation) have cesophageal pain may be indistinguishable from also shown only small and variable values for the that of myocardial infarction. The second type of Pco2 difference, but in these dogs the ventilation pain is pleural pain, which is usually only present in the embolized lung was reduced by only a if subpleural infarction occurs. Cyanosis is com- small amount. The end-tidal, arterial Pco2 dif- mon and is due partly to the low cardiac output ference will be considerably reduced if, instead of (stagnant anoxia) and partly to the development the embolus completely blocking a lung or a region of right to left shunts in the lungs (Robin and of a lung, it only partly obstructs it and still allows others, I960; Stein, Forkner, Robin and Wessler, some blood to flow to the region. Oxygen uptake I96I). Circulatory collapse may occur when the measurements in dogs with large pulmonary pulmonary circulation is severely obstructed. The emboli suggest that some blood does get past a cause of the shock is probably a combination of Postgrad Med J: first published as 10.1136/pgmj.38.435.13 on 1 January 1962. Downloaded from

3fanuary I962 J9MARSHALL: Pulmonary Emibolism and Thrombosis low Lardiac output and reflex peripheral vaso- of subintimal fibrosis containing blood vessels. In dilatation. Transient loss of consciousness is not these dogs there was no evidence of secondary uncommon. Overloading of the right heart mani- thrombosis around the original embolus, but fests itself by increased jugular venous pressure secondary thrombosis, proximal and distal to the and electrocardiographic changes (Krause and embolus, has been reported to be common in man Silverblatt, 1955). The increased venous pressure (Schoenmackers, 1958) and to be responsible for is often an indication of the increased filling extensive infarction in some cases. pressure required to maintain the cardiac output Todd (I959) has demonstrated by histochemical against an increased resistance rather than a sign methods that fibrinolysin activators are present in of heart failure. Electrocardiographic changes are the intima of the pulmonary artery and also in by no means always present even in clinically veins, but are absent from systemic arteries. This apparent embolism. Similarly, although radio- is of interest in view of the tendency of pulmonary graphic changes of dilatation of the pulmonary thrombosis to occur in patients with pulmonary artery and increased translucency of the affected hypertension. The thrombosis may occur, not lung have been described, X-ray changes are un- because of intimal damage to the arteries, but common in embolism uncomplicated by infarction. because the pulmonary arteries resemble systemic When infarction occurs linear or wedge-shaped vessels and do not contain fibrinolysin activator. opacities may be seen on the X-ray. Heemoptysis is not a characteristic sign of uncomplicated em- Incidence bolism and only occurs if infarction is present; Pulmonary embolism has been reported to occur even so, only about 50% of cases of'infarction in 5 to I4% of all autopsies (Parker and Smith, suffer from hemoptysis. 1958) and at the Radcliffe Infirmary at Oxford during the year 1957 pulmonary embolism was the The Fate of Pulmonary Emboli cause of death in 70 (8%) of the 838 autopsies. Patients who are known to have had a pullmonary Such post-mortem evidence of infarction is prob- embolism often show no clinical abnormality after- ably an underestimate of the total cases because by copyright. wards and cardiac and respiratory tests may be small emboli may be missed and old emboli may normal when carried out a few months or years be absorbed or may only be evident as changes afterwards (Duner, Pernow and Rigner, 1960). in the vessel wall (Harrison, 1948; Allison and Many of these emboli, to judge from the symptoms others, 1960). and signs at the time of impaction, must have been Pulmonary infarction commonly accompanies fairly massive. On the other hand, some patients, embolism in old patients and in patients with con- particularly those with repeated emboli, may gestive heart failure or pulmonary infections develop cor pulmonale afterwards (Ehrner, Garlind (Roach and Laufman, I954), but it is unusual in and Linderholm, 1959). Most emboli are com- young patients with no complicating factors. It posed of blood thrombus, which is susceptible to is usually stated that pulmonary infarction does http://pmj.bmj.com/ attack by fibrinolytic mech'anisms. The recovery not occur in normal lung in which only the pul- of function after to-day-old autologous thrombi monary artery is obstructed; the bronchial artery 'had been impacted in the pulmonary arteries of supply is sufficient to maintain the nutrition of the dogs has been followed by serial bronchospiro- lung. The lung can withstand ischaemia extremely metry and measurement of diffusing capacity well. During replantation of dog lungs the lung (Marshall and others, in preparation). The may be completely ischaemic and unventilated for recovery of oxygen uptake by the embolized lung over two hours without apparent damage. Ligature on September 24, 2021 by guest. Protected started a few minutes after the embolus had im- of the pulmonary veins in dogs causes hiemorrhagic 'pacted, was rapid over the first few days and then cedema of the affected lung. Venous thrombosis slowed down so that when the dogs were sacrificed may be essential for infarction to develop. In a at 14 days the lung had recovered about 8o0% of its lung affected by cedema or with increased secre- original function. tions bronchial obstruction may easily occur, cutting off the direct supply of oxygen to the lung Pathology parenchyma and facilitating tissue necrosis. Blood clot emboli in dogs become adherent to the vessel wall and overgrown by ' endothelium ' Prevention (Allison, Dunnill and Marshall, I960). Vasculariza- Most cases of pulmonary embolism occur in tion of the vessel wall increases at the point of patients with congestive heart failure or after attachment of the clot and capillaries and fibro- operations and methods of prevention should be blasts grow into the clot. The clot is rapidly directed towards preventing peripheral venous absorbed so that in dogs examined at 2I days thrombosis in these patients. Although pulmonary there was no recognizable clot, but only focal areas embolism typically occurs about io days after Postgrad Med J: first published as 10.1136/pgmj.38.435.13 on 1 January 1962. Downloaded from

I6 POSTGRADUATE MEDICAL JOURNAL 7anuarv I962 operation, the pathological process leading to this Fat embolism may occur following fracture of event probably starts on the operating table. Con- bones or damage to fatty tissues. This may cause siderable venous stasis has been demonstrated in symptoms similar to those of blood-clot embolism, the legs of patients lying horizontally on an oper- but, in addition, some of the fat may be squeezed ating table (McLachlin, McLachlin, Jory and through the pulmonary capillaries and cause Rawling, I960); radio opaque dye injected into the cerebral embolism. Air in the pulmonary circula- veins may remain for up to one hour in the valve tion causes obstruction because the surface tension sinuses of the deep veins of the legs. This stasis, of the air-blood interface is too great to allow the often accompanied by prolonged pressure on soft air to be forced through the pulmonary capillaries tissues and by accentuation of the clotting factors by the right heart. Emboli of tumour cells may be as a result of the operative trauma, may well fairly large, as when a fragment breaks off a hyper- initiate thrombus formation which extends during nephroma growing into the inferior vena cava, or the quiet post-operative period. Measures to there may be multiple small emboli from a chorion- prevent thrombosis should be taken throughout epithelioma. Embolism by chorionepithelioma the period that the patient is anmesthetized; early cells usually occurs insidiously and is one of the active movements after operation may be too late. common causes of ' primary' pulmonary hyper- Raising the legs to 150 above horizontal has been tension. Tumour cell emboli are not susceptible shown to reduce venous stasis during operation to attack by fibrinolysins and are probably more and this should be combined with the avoidance effective in causing permanent blockage of the of pressure on the leg muscles. pulmonary vessels than are small blood-clot The prevention of thrombosis in patients with emboli. congestive heart failure is more difficult. The old- fashioned ' donkey ' which lay under the knees of Treatment an orthopnceic patient and whit h prevented venous The effects of pulmonary embolism, for which return from the legs has pass-ed out of use. In treatment may be required, can be considered

patients who are confined to bed active or passive under three headings: by copyright. movements of the legs combined with periodical elevation will help'to prevent stasis in the'legs, but (i) Obstruction to the Circulation can do little to prevent thrombosis in the pelvic The right ventricle may,fail as a result of the veins. Anticoagulants are of some value in avoid- increased resistance against which it has to work ing thrombosis, but art by no means a sure pre- and the reduced cardiac output may be unable to ventative. Once thronmbosis has occurred treat- maintain the systemic blQod pressure, causing ment is directed towards keeping the thrombi in stagnant anoxia and clinical ' shock '. Morphine is the peripheral veins until they have become usually of considerable value in countering the organized or absorbed, or towards preventing their patient's anxiety-in spite of its action in lowering peripheral vascular resistance. If the systemichttp://pmj.bmj.com/ passage to the heart and lungs. Immobilization of pressure is very low 1-noradrenaline may be the legs is widely practised, but if this is to be required to increase the peripheral resistance. This effective it should probably be maintained rigor- drug has only a weak action on the pulmonary ously for several weeks. Ligature of the inferior circulation, so that its use in pulmonary embolism vena cava is now seldom used because of the leg is not contraindicated. Digitalis is reported to be cedema which it often causes, but ligature of the of little value unless the heart was in failure before common iliac veins allows collateral venous cir- the onset of embolization. Oxygen administration culation to develop and is used in selected cases. increases the oxygen-carrying power of the blood on September 24, 2021 by guest. Protected Suitable patients for vein ligation are those with by virtue of the dissolved oxygen and is also recurrent emboli and obvious thrombosis in the valuable because the diffusing capacity of the legs. lungs may be reduced in patients with heart disease Embolism by other Substances than Blood or pulmonary cedema. An apparatus should be Clots or Thrombi used which gives as high as possible a concentra- Substances other than blood clot which may tion of oxygen. Venesection is contraindicated cause pulmonary embolism are fat, amniotic because it results in further lowering of the cardiac fluid, air and tumour cells. Damage to the placenta output. may allow the passage into the maternal circulation' (2) Possible Reflex Effects of amniotic fluid containing desquamated cells, The existence of reflex vasoconstriction as a meconium, hairs and other debris. This results in result of pulmonary embolism has not been sub-' severe shock which may be partly an anaphylactic stantiated and neither in exDerimental animals nor reaction to the amniotic fluid in addition to the in human cases have ' vasodilators ' been found to effect of blockage of the pulmonary capillaries. be of benefit. Atropine and papaverine have been Postgrad Med J: first published as 10.1136/pgmj.38.435.13 on 1 January 1962. Downloaded from January I962 MARSHALL: Pulmonary Embolism and Thrombosis '7 the drugs most commonly tried for this purpose. (3) Superadded Thrombosis Reflex bronchoconstriction may occur as a result Thrombosis may occur in the branches of the of embolism. Broncho-constriction of the em- pulmonary artery affected by the embolus (Schoen- bolized lung is beneficial in reducing the ventila- mackers, 1958) and lead to infarction. The throm- tion of the non-perfused lung (Severinghaus and botic process in the peripheral veins also requires others, I96I), but if it is generalized it may require treatment and favours the administration of anti- treatment and aminophylline is probably the most coagulants in all cases. Hemorrhagic infarction is effective drug. not a contraindication.

REFERENCES ALLISON, P. R., DUNNILL, M. S., and MARSHALL, R. (I960): Pulmonary Embolism, Thorax, 15, 273. BROFMAN, B. L., CHARMS, B. L., KOHN, P. M., ELDER, J., NEWMAN, R., and RiziKA, M. (1957): Unilateral Pulmonary Artery Occlusion in Man, J3. thorac. Surg., 34, 2o6. CoMROE, J. H., VAN LINGEN, B., STROUD, R. C., and RONCORONI, A. (1953): Reflex and Direct Cardiopulmonary Effects of s-OH-Tryptamine (Serotonin): Their Possible Role in Pulmonary Embolism and Coronary Thombosis, Amer. J. Physiol., 173, 379. DY, R. (1957): The Autonomic Nervous System and its Relationship to Some Forms of Heart and Lung Disease, /Brit. med. J7., ii, 249. p :R, H., PERNOW, B., and RIGNER, K. G. (I960): The Prognosis of Pulmonary Embolism: A Medical and Physio- logical Follow-up Examination of Patients Treated at the Department of Internal Medicine and Surgery, Karo- J linska Sjukhuset, in I952-58, Acta mod. scand., I68, 38I. <;HRNER, L., GARLIND, T., and LINDERHOLM, H. (1959): Chronic Cor Pulmonale Following Thromboembolism: A Clinical and Pathophysiological Study of Three Cases, Ibid., I64, 279. HARRISON, C. V. (1948): Experimental Pulmonary Arteriosclerosis, J. Path. Bact., 60, 289. JULIAN, D. G., TRAvis, D. M., ROBIN, E. D., and CRUMP, C. H. (I960): Effect of Pulmonary Artery Occlusion Upon End-tidal CO2 Tension, J. appl. Physiol., 15, 87. {RAUSE, S., and SILVERBLATT, M. (1955): Pulmonary Embolism: A Review with Special Emphasis on Clinical and Electro-cardiographic Diagnosis, Arch. intern. Med., 96, I9. McEvoy, R. K., HARDER, R. A., and DALE, W. A. (1958): Respiratory and Cardiovascular Phenomena Associated with Pulmonary Embolism, Surg. Gynec. Ostet., Io6, 271. by copyright. McLACHLiN, A. D., McLACHLIN, A. J., JoRY, T. A., and RAWLING, E. G. (I960): Venous Stasis in the Lower Ex- tremities, Ann. S;trg., 152, 678. MARSHALL, R., ALLISON, P. R., BOSMAN, A. R., SABISTON, D. C., and DUNNILL, M. S.: A Study of the Immediate j' and Later Effects of Pulmonary Embolism by Large Blood Thrombi in Dogs (in prepaiation). NIDEN, A. H., and AvIADO, D. M. (1956): Effects of Pulmonary Embolism on the Pulmonary Circulation with Special Reference to Arteriovenous Shunts in the Lung, Circulat. Res., 4, 67. PARKER, B. M., and SMITH, J. R. (I958): Pulmonary Embolism and Infarction, Amer. J. Med., 24, 402. ROACH, H. D., and LAUFMAN, H. (i955): Relationship Between Pulmonary Embolism and Pulmonary Infarction: and Experimental Study, Ann. S'rg., 142, 82. ROBIN, E. D., FORKNER, C. E., BROMBERG, P. A,, CROTEAU, J. R., and TRAvIS, D. M. (I960): Alveolar Gas Exchange in Clinical Pulmonary Embolism, Newv Engl. J. Med., 262, 283. SANDERS, R., WAALKES, J. P., GILBERT, J. W., and TERRY, L. L. (I959): Serotonin (5-Hydroxytryptamine) and Pul- http://pmj.bmj.com/ monary Thromboembolism, Ssrg. Gynec. O.stet., 109, 455. SCHOENMACKERS, J. (1958): Zur Pathologie der Lungenarterienembolie, Dtsch. med. Wschr., 83, 115. SEVERINGHAUS, J. W., and STUPFEL, M. (T957): Alveolar Dead Space as an Index of Distribution of Blood Flow in Pulmonary Capillaries, J. appl. Physiol., 10, 335. -, SwENSON, E. W., FINLEY, T. N., LATEGOLA, M. T., and WILLIAMS, J. (196I): Unilateral Hypoventilation Pro- duced in Dogs by Occlusing One Pulmonary Artery, Ibid., I6, 53. STEIN, M., FORKNER, C. E., ROBIN, E. D., and WESSLER, 5&. (I96I): Gas Exchange After Autologous Pulmonary Em- bolism in Dogs, rbid., I6, 488.

TODD, A. S. (i959): The Histological Localization of Fibrinolysin Activator, J. Path. Bact., 78, 281. on September 24, 2021 by guest. Protected WILLIAMS, M. H. (1956): Mechanical versus Reflex Effects of Diffuse Pulmonary Embolism in Anlesthetized Dogs, Circulat. Res., 4, 325.