294 COLLAPSE THERAPY (MEDICAL) IN PULMONARY TUBERCULOSIS By F. H. SCADDING, M.D., M.R.C.P. Assistant Physician, Middlesex Hospital Whatever the exact mechanism involved in the various measures embraced by the term ' collapse therapy' for pulmonary tuberculosis, a feature common to them all is the reduction in volume of the diseased portion of lung. There is ample evidence that the resultant relaxation is beneficial and conducive to healing. Yet, even before the advent of chemotherapy and the progress in thoracic surgery which the past decade has wit- nessed, the mainstay of minor collapse therapy, the pneumothorax, was receiving its full share of criticism. Unfavourable general conclusions, how- ever, were too often drawn from what we should now regard as its misuse. Many past failures can today be understood as a result of our greater knowledge of the pathology of the disease, especi- ally as regards the important role played by the bronchi. The essential factors for success of any treatment are a proper selection of cases and a clear recognition of the limitations of the par- ticular method employed. All are agreed that cavitated disease is the main indication for some form of collapse therapy. In such cases the risk of further spread of the disease by bronchial aspiration to the opposite lung or to other parts of the same lung is ever present, and the ultimate prognosis of patients with persistent cavitation is very poor indeed. Bed rest alone, if strict and prolonged, will cause some cavities to close. If, at the same time, a position in bed is adopted which places the cavity in the most dependent position, as suggested by Dilwyn Thomas (1950), even large cavities can sometimes be persuaded to close or shrink considerably in size. Without additional mechanical control, however, the ten- dency for them to re-open when the patient becomes ambulant is very great. Effective anti- bacterial treatment of recent years raised hopes that collapse procedures might soon become unnecessary. Such hopes, however, have not yet been realized. Recent pulmonary cavities will often appear to close on chemotherapy, but if no additional measures are adopted, the relapse rate is high when the treatment is stopped. For disease which has progressed to cavitation, there- fore, chemotherapy must be integrated with other forms of treatment and not replace them. It has done much to make them safer by rendering the disease quiescent before collapse therapy, and has also reduced the incidence of the more serious complications; to a lesser extent it has brought unsuitable cases within reach of collapse therapy. Whether to apply collapse measures for un- cavitated tuberculous infiltration remains a dif- ficult problem, the solution of which depends largely on personal judgment and experience. Assuming there is evidence of activity, anti- bacterial treatment should be started without delay together with bed rest, preferably under close observation in hospital. If considerable radiological clearing does not result after six weeks of such a regime, or should there be deterioration, some form of collapse should be considered. Many factors other than the X-ray appearance and clinical condition weigh in this decision, e.g. the young woman with dependent family, a bad family history of tuberculosis, racial predisposition-such features would rightly in- fluence the decision for active treatiflent in cases of doubt. Phrenic Interruption Various operations on the phrenic nerve from simple freezing with ethyl chloride to avulsion were frequently practised in the 193os, but today the only one in regular use is phrenic crush. It produces only a very moderate degree of pul-, monary relaxation and, although not quite obsolete, as a sole measure of collapse therapy it has a very limited sphere of application. To have any measurable effect the diaphragm should be mobile prior to the operation, and the lung should not be widely adherent to the chest wall. Pre- liminary fluoroscopy will give some idea of the expected rise of the diaphragm, which on the average will be 2 to 4 cm. The effect of a single crush usually lasts from six to nine months, but it should be borne in mind that about 50 per cent. copyright. on June 19, 2020 by guest. Protected by http://pmj.bmj.com/ Postgrad Med J: first published as 10.1136/pgmj.30.344.294 on 1 June 1954. Downloaded from
By F. H. SCADDING, M.D., M.R.C.P.Assistant Physician, Middlesex Hospital
Whatever the exact mechanism involved in thevarious measures embraced by the term ' collapsetherapy' for pulmonary tuberculosis, a featurecommon to them all is the reduction in volume ofthe diseased portion of lung. There is ampleevidence that the resultant relaxation is beneficialand conducive to healing. Yet, even before theadvent of chemotherapy and the progress inthoracic surgery which the past decade has wit-nessed, the mainstay of minor collapse therapy,the pneumothorax, was receiving its full share ofcriticism. Unfavourable general conclusions, how-ever, were too often drawn from what we shouldnow regard as its misuse. Many past failures cantoday be understood as a result of our greaterknowledge of the pathology of the disease, especi-ally as regards the important role played by thebronchi. The essential factors for success of anytreatment are a proper selection of cases and aclear recognition of the limitations of the par-ticular method employed. All are agreed thatcavitated disease is the main indication for someform of collapse therapy. In such cases the riskof further spread of the disease by bronchialaspiration to the opposite lung or to other partsof the same lung is ever present, and the ultimateprognosis of patients with persistent cavitation isvery poor indeed. Bed rest alone, if strict andprolonged, will cause some cavities to close. If,at the same time, a position in bed is adoptedwhich places the cavity in the most dependentposition, as suggested by Dilwyn Thomas (1950),even large cavities can sometimes be persuaded toclose or shrink considerably in size. Withoutadditional mechanical control, however, the ten-dency for them to re-open when the patientbecomes ambulant is very great. Effective anti-bacterial treatment of recent years raised hopesthat collapse procedures might soon becomeunnecessary. Such hopes, however, have not yetbeen realized. Recent pulmonary cavities willoften appear to close on chemotherapy, but if noadditional measures are adopted, the relapse rateis high when the treatment is stopped. For
disease which has progressed to cavitation, there-fore, chemotherapy must be integrated with otherforms of treatment and not replace them. It hasdone much to make them safer by rendering thedisease quiescent before collapse therapy, andhas also reduced the incidence of the more seriouscomplications; to a lesser extent it has broughtunsuitable cases within reach of collapse therapy.Whether to apply collapse measures for un-
cavitated tuberculous infiltration remains a dif-ficult problem, the solution of which dependslargely on personal judgment and experience.Assuming there is evidence of activity, anti-bacterial treatment should be started withoutdelay together with bed rest, preferably underclose observation in hospital. If considerableradiological clearing does not result after sixweeks of such a regime, or should there bedeterioration, some form of collapse should beconsidered. Many factors other than the X-rayappearance and clinical condition weigh in thisdecision, e.g. the young woman with dependentfamily, a bad family history of tuberculosis, racialpredisposition-such features would rightly in-fluence the decision for active treatiflent in casesof doubt.
Phrenic InterruptionVarious operations on the phrenic nerve from
simple freezing with ethyl chloride to avulsionwere frequently practised in the 193os, but todaythe only one in regular use is phrenic crush. Itproduces only a very moderate degree of pul-,monary relaxation and, although not quiteobsolete, as a sole measure of collapse therapy ithas a very limited sphere of application. To haveany measurable effect the diaphragm should bemobile prior to the operation, and the lung shouldnot be widely adherent to the chest wall. Pre-liminary fluoroscopy will give some idea of theexpected rise of the diaphragm, which on theaverage will be 2 to 4 cm. The effect of a singlecrush usually lasts from six to nine months, butit should be borne in mind that about 50 per cent.
SCADDING: Collapse Therapy (Medical) in Pulmonary Tuberculosis
obtain only a partial recovery of function, andmore than io per cent. do not recover at all.Regeneration proceeds more slowly with increasingage, so this operation should seldom be used inpatients over fifty. It is quite useless for extensivefibrous disease or thick walled cavities, and shouldbe avoided in patients likely to require majorsurgery in view of the higher incidence of post-operative complications. It still has advocates forits trial in patients with limited exudative diseasewhose response to rest and chemotherapy is slow.Its main role, however, is a supplementary one,either as an adjunct to a pneumopsritoneum, orto improve an inadequate pneumothorax in which,because of apical and basal adhesions, the affectedlung is only relaxed in the lateral plane. Atoperations for lung resection it can be used todiminish the capacity of the hemithorax and, atthe termination of a pneumothorax, particularlyif the lung is slow to re-expand, a phrenic crushwill enable a more gradual adjustment of theexpanding lung and help to lessen any tendency tomediastinal displacement. If used for this pur-pose the operation should be done before the lowerlobe makes contact with the chest wall.
Artificial PneumothoraxThere are several reasons for the waning
popularity of pneumothorax treatment of pul-monary tuberculosis over the past ten years.General conclusions drawn from its use in thewrong types of case have in part been responsible.Other important factors are the introduction ofa simple and safe alternative in the form ofpneumoperitoneum, the availability of effectivechemotherapy and the great strides made bythoracic surgery. Nevertheless, a good pneumo-thorax is still the most effective minor collapseprocedure, since it provides concentric relaxationand is applicable to disease in any part of thelung. It must be realized, however, that manyforms of pulmonary tuberculosis are unsuitablefor A.P. treatment, either because it is ineffectiveor positively dangerous. Experience has taught usto recognize the following contraindications:
(i) Tuberculous Pneumonias. Included in thisgroup are also the more florid forms of the diseasewith extensive infiltration and clinical evidence ofsevere toxaemia. In such cases chemotherapy,possibly assisted by a pneumoperitoneum, shouldbe- the first line of treatment and may in someinstances render subsequent A.P. treatment safe,but-the induction of a pneumothorax in the earlystages- carries a great risk of pleural effusion andempyema.
(ii) Tuberculoma. The solid round focus mayrepresent- a blocked cavity, a circumscribed areaof caseous pneumonia or a cold abscess, of a
bronchus. An A.P. has no material effect on suchlesions and should therefore not be used.
(iii) Chronic Disease with Extensive Fibrosis ormuch Destruction of Lung Tissue. An effectiveA.P. is seldom obtainable because of indivisiblepleural adhesions and, in general, such casesdemand a more permanent form of relaxation,or resection.
(iv) Large Peripheral Cavities. These are almostalways adherent to the chest wall and may receivepart of their blood supply through such adhesions.If these should be severed the wall may necroseand allow the cavity to rupture into the pleuralspace with disastrous consequences.
(v) Tuberculous Disease of Major Bronchi.Tuberculous bronchitis should be suspected wherethe X-ray shows evidence of atelectasis or a' tension' type of cavity. The narrowing of abronchus caused by tuberculous inflammation ofits wall may result in a check-valve mechanismwhereby air may enter a cavity under conditionsof raised intra-bronchial pressure, as duringcoughing, but be unable to escape during expira-tion. Such a cavity tends to be spherical, thin-walled and often contains a fluid level in view ofthe -impaired bronchial drainage. PermanentClosure of the affected bronchus will result incavity closure, and this can sometimes be achievedby an A.P. Unfortunately, however, the check-valve effect may only be aggravated with a con-sequent rapid increase in size of the cavity andpossible rupture. It is wiser, therefore, to avoida pneumothorax in such cases unless the tensionelement in the cavity can first be overcome bychemotherapy.The exclusion of all the above groups decidedly
narrows the field for A.P. treatment in pulmonarytuberculosis. Nevertheless, there remains a con-siderable number of patients with recent scatteredinfiltration, often with cavities of less than 4 cm.diameter, and mainly or entirely unilateral dis-tribution. For these a pneumothorax is mostuseful. Its induction, however, should be regardedas an exploratory procedure and it should beabandoned if it is not effective or cannot be madeeffective by adhesion section. If feasible, alladhesions preventing relaxation of the diseasedarea should be divided within the first few weeksof induction. Considerable experience is requiredto decide on the safety and extent of this operation,as the most serious complications of A.P. treatmentoften date from it.The importance of undivided adhesions is still
a matter of controversy. There is, however,general agreement that a contra-selective pneumo-thorax, or one under which a cavity persists,should not be-maintained. They are fraught withdanger and, as Rafferty (I947) has shown, the
eune I954 295copyright.
on June 19, 2020 by guest. Protected by
http://pmj.bm
j.com/
Postgrad M
ed J: first published as 10.1136/pgmj.30.344.294 on 1 June 1954. D
prognosis for such patients is worse than if theyhad no active treatment. Some authorities con-sider that only those A.P.s should be maintainedin which the lung is completely free. A moremoderate view was expressed by Scadding et al.(I95I), who found that adherence of the upperlobe to the superior mediastinum did not neces-sarily preclude success nor carry any great risk.Foster-Carter et al. (1952) reviewed 457 patientstreated by artificial pneumothorax at Bromptonand its sanatorium five years after their A.P.s hadbeen given up. The survival rate in patients witha completely free lung was 9I per cent. and, inthe ' adherent satisfactory' group of 242 patients,where adhesions to the diseased area were presenitbut did not prevent cavity closure within a year,the survival rate was 94 per cent. Furthermore,the incidence of persistent pleural effusions wasthe same in the two groups. They concluded thatthe key to successful pneumothorax treatment wasclosure of cavities and not the presence of in-divisible adhesions. With proper selection mostcavities will close within two months, but a fewmay take several months. In the absence ofcomplications, therefore, an imperfect pneumo-thorax need not be abandoned in favour of surgerybefore its merits have been fairly assessed aftera trial period of adequate and frequent refills.
Pleural Effusion and ' Frozen ' LungA small collection of fluid in the costo-phrenic
angle is frequently observed during treatment bypneumothorax. Remaining below diaphragmlevel, it seldom interferes with the course of theA.P. and requires no treatment other than obser-vation. Occasionally it appears to initiate pre-mature obliteration of the pleural space, but isotherwise harmless. The larger, cumulativeeffusions, however, are of more serious import.If they do not absorb within a few weeks of treat-ment by extra rest, aspirations and lavage, it isprobably wise to abandon the A.P., even thoughmore drastic surgical treatment may then have tobe faced. Evidence is accumulating to show thatinitial chemotherapy reduces the incidence ofthese effusions and of tuberculous empyemafollowing adhesion section.The factors which predispose to the complica-
tion of ' frozen ' lung are the persistent pleuraleffusion, extensive pulmonary fibrosis and per-manent atelectasis. It can often be anticipatedand, therefore, should be largely preventable byavoiding A.P. treatment in cases with bronchialdisease or a considerable fibrotic element, and bythe early abandonment of those pneumothoracesin which progressive pleural thickening or effusionoccurs. Once the condition is fully developedthe loss of pulmonary function can seldom be
regained even by decortication of the affectedlung (Wright et al., 1949).
Duration and TerminationThe aim of treatment is to continue the pneumo-
thorax until the disease is healed, but to give it upwhile the lung is still capable of re-expansion.There can be no fixed rule about its durationsince such variable factors as the character andextent of the original lesion, age and sex of thepatient, type of pneumothorax and its complica-tions have to be considered in each case. Ingeneral, a period of three years will be required,with extension to five years if cavities were presentinitially. It may well be that chemotherapy inthe early months of treatment may shorten thetime necessary for healing, but there is not yetsufficient evidence to warrant any drastic reductionin the period of collapse. When the decision todiscontinue the A.P. has been taken, a progressivereduction in either the frequency or the amountof refills is preferable to an abrupt cessation. Themore gradual adjustment minimises the patient'sdiscomfort and diminishes the risk of late effusiomor tearing of an indurated visceral pleura. Tomo-graphy should be done before the lung reachesand adheres to the chest wall as, in the few whosecavities reopen, the pneumothorax may be regainedif further minor collapse therapy is thought likelyto succeed.
PneumoperitoneumThe introduction of air into the peritoneal
cavity as a treatment for bilateral pulmonarytuberculosis was suggested by Vajda in 1933.Banyai (I934), combining it with phrenic paralysis,explored its possibilities as a definitive form ofcollapse therapy, with favourable results, andsince 1943 it has been widely practised in thiscountry. It came as a welcome alternative to thepneumothorax and, indeed, it has certain advan-tages over it in that the technique is easy to acquire,serious complications are rare, it can be given upand re-induced at will, it can safely be used inthe patient's own home and, above all, it is freefrom serious pleural and pulmonary complications.Furthermore, it can be applied to the acute formsof the disease, bilateral cases, and those in whicha satisfactory A.P. cannot be obtained. In somechest clinics it has almost replaced artificialpneumothorax as the first choice when collapsetherapy is indicated, and a large volume of medicalliterature testifies to its value. Most of theenthusiastic reports, however, are concerned onlywith early results.The effect of pneumoperitoneum, generally
speaking, is proportional to the degree of diaphrag-matic elevation obtained, and little benefit can be
296 )Iune 1954copyright.
on June 19, 2020 by guest. Protected by
http://pmj.bm
j.com/
Postgrad M
ed J: first published as 10.1136/pgmj.30.344.294 on 1 June 1954. D
June 1954 SCADDING: Collapse Therapy (Medical) in Pulmonary Tubercuiosts 297
.... ...
L. 2 :'
FIG. I (a).-Tuberculous pneumonia. Adolescent male,acutely ill with high fever and severe toxaemia;signs of consolidation right lower lobe and infil-tration left mid-zone. Sputum positive.
.. ...
L.. 1.4:.
FIG. i (c).-After seven months chemotherapy com-bined with right phrenic and P.P. The diaphragmhas recovered. Patient symptom-free, havinggained over a stone in weight and beginning to getup. One gastric lavage culture positive.
... ...e
:: :::aNi .}::!.:, o.o., :
b-. :::...:
FIG. I (b).-After three weeks treatment with dailystreptomycin and P.A.S. Some clinical improve-ment but lesion rapidly breaking down to formseveral cavities. Right phrenic crush and P.P.instituted.
R.M k 4 f1
FIG. I (d).-Tomogram showing persistent cavitation inright lower lobe. Two shrunken, bronchiectaticsegments subsequently resected from right lowerlobe without incident. Patient well and workingto date.
expected if the diaphragm is fixed, or if the pleurais widely adherent. It is quite common for theparalyzed diaphragm to rise to the level of thethird rib anteriorly or even higher. Under suchcircumstances it can control lesions in the upperparts of the lung as effectively as those in thelower, provided they are of a type likely torespond (Anderson and Winn, I945; Keers,I948). The following types of disease respondwell:
(i) Acute Exudative Lesions. In cases of tuber-culous pneumonia the combination of chemo-therapy with a P.P. and phrenic is often life-saving. It may even be sufficient to bring quies-cence and eventual healing, but more often servesto tide the patient over a critical phase of hisillness and enables more definitive treatment ata later date. Whilst it has virtually no effect uponchronic cavities in the upper lobes, it may offerconsiderable help in controlling acute spreadsfrom such cavities in either lung, and so renderthe patient acceptable for surgical treatment.This use of pneumoperitoneum as an interimmeasure probably represents its most importantsphere of application in the management ofpulmonary tuberculosis.A good example is seen in Figs. i (a) to (d).(ii) Predominantly Exudative Disease with Recent
Cavitation, in which an A.P. has Failed or isContraindicated. In such conditions a P.P. maybe used with safety and is certainly preferable tomaintaining an inadequate pneumothorax.
(iii) Lower Lobe Disease. A common site forpulmonary tuberculosis is the apical segment ofthe lower lobe, and in this situation it has oftenprogressed to cavitation by the time it is diagnosed.Crofton (I949) showed that a good pneumothoraxwould close two-thirds of such cavities, and thisis probably the treatment of choice, provided thedraining bronchus can be shown at bronchoscopynot to be stenosed. If, however, a good A.P. isunobtainable then the effect of a P.P. and phrenicis worth trying.
(iv) Severe Haemoptysis. It is of interest thatits value in this connection was noted by Banyaiin 193I when, inducing an A.P. for the samepurpose, air was accidently injected into theperitoneal cavity. Most of the haemoptyses inthe course of pulmonary tuberculosis cease on theirown or respond to simple measures. Successfularrest of the more severe cases by induction ofa P.P. has been reported, and it has the advantageover pneumothorax that the site of the haemor-rhage need not be known.
ComplicationsAlthough the safety of the procedure has been
stressed, more than 40 complications have been
described, varying in severity from vague indiges-tion to sudden death. Perforation of abdominalviscera has often been reported but causes sur-prisingly little anxiety. The more serious com-plications include air embolism, intraperitonealhaemorrhage, spontaneous pneumothorax, medias-tinal emphysema, torsion of the spleen and tuber-culous peritonitis. Luckily they are all uncommon.Congestive heart failure may be precipitated inpatients with a poor cardiac reserve, and anincreased incidence of acute appendicitis has beennoted (Rilance and Warring, I944).
Duration of TreatmentWhere a P.P. is employed as an interim measure
it will usually be required for a few months, butthe time will vary according to the circumstancesof each case. The optimum time for surgery,and whether the P.P. should be abandoned ormaintained through the post-operative period, arematters for discussion between the physician andsurgeon concerned. If the diaphragm is stillparalyzed some surgeons prefer to keep the P.P.going until after the operation in the belief thatthe risk of post-operative atelectasis is therebylessened.As with pneumothorax, there can be no set
time for which a P.P. should be maintained whenit is the only means of collapse adopted. It isunlikely to prove effective in a shorter time thanwould a pneumothorax and a minimum period ofthree years is, therefore, advised. During thistime it may be necessary to re-crush the phrenicnerve on more than one occasion, but it is oftenobserved that the diaphragm remains high inposition even after recovering its function. Atthe termination of the treatment no ill effectsresult from the abrupt cessation of refills, and theair has usually absorbed in 6 to I2 weeks.
In the absence of a specific remedy it takes manyyears to assess the value of any treatment for adisease with such a variable natural course aspulmonary tuberculosis. The simplicity andrelative harmlessness of pneumoperitoneum treat-ment have encouraged its trial in all varieties ofthe disease, but the temporary relief of symptomnsand psychological benefit sometimes afforded, donot justify its continued use in unsuitable cases.The late results are only just appearing and theyare disappointing in advanced cases (Morris et al.,1952). Edge (I953) confirms this, and reportsalso a high relapse rate (g out of 25) soon afterabandonment in patients with anything more thanearly lesions. It seems likely that such results,together with further progress in other forms oftreatment, will cause pneumoperitoneum- to beused with greater discrimination.Continued on page 304
-.)une ^I-9g4298copyright.
on June 19, 2020 by guest. Protected by
http://pmj.bm
j.com/
Postgrad M
ed J: first published as 10.1136/pgmj.30.344.294 on 1 June 1954. D
length of the fourth rib is resected. The extra-pleural plane is entered through the rib bed andthe apex of the lung is stripped off the endo-thoracic fascia to a similar extent to that for thetwo preceding operations. When absolutehaemostasis has been achieved, and this may bedifficult, the wound is closed with particular careto ensure that the intercostal layer is air tight. Thecollapse is then maintained by filling the spacewith air at suitable intervals.
Results ofCollapse OperationAs the extent and the type of tuberculous
lesions which are subjected to surgery vary soenormously, the statistical results in any seriesmust depend to a very large degree on the type ofcase operated upon. For instance, if a high pro-portion of the patients are suffering from advancedbilateral disease the mortality rate may well be ashigh as io per cent., whereas if the majority of theoperations performed were for ' minimal' apicallesions without cavitation, no mortality greaterthan i per cent. would be acceptable.The overall mortality for thoracoplasty is
approximately 3 per cent., and it is the same forplombage operations. In published series (PriceThomas, I952; Sellors, 1947; Laird, I953) it isin the same order.
In terms of sputum conversion the results of mostcollapse operations give figures ranging around8o per cent. successes, but before conversion canbe accepted at least three negative sputum orgastric lavage cultures are required, as the morethorough the search for bacilli the more often theyare found. Short of post-mortem evidence cavityclosure is almost impossible to prove. In terms ofreturn to their previous occupation Price Thomasstates that in his experience over 8o per cent. ofthe patients surviving operation are back in theiroriginal job. Early post-operative morbidity ishigher after thoracoplasty than after plombageoperations, but late tuberculous infection of theextra-periosteal space is the bugbear of the latteroperation in which foreign bodies are left in situ,and it still occurs sufficiently often for manysurgeons to doubt the justification of the pro-
cedure. This criticism also applies to extra-pleural pneumothorax, in which late tuberculousinfection is not uncommon; this operation has theadded disadvantage that refills are required weeklyand that despite this inconvenience the lungseldom re-expands when refills are discontinuedand so its basic purpose is not achieved.
Summary.i. Surgical collapse therapy has a large place in
the treatment of pulmonary tuberculosis.2. The decision as to the extent and the type of
collapse should be based on as complete as possibleanatomical and pathological diagnosis of the disease.
3. Thoracoplasty has the disadvantage that ithas to be done in more than one stage, but thelong-term results are probably better than those inany other collapse operation.
4. Extra periosteal plombage has the advantageof being a one-stage procedure, but the dis-advantages inherent in any operation in whichforeign material is left in situ.
5. Extra-pleural pneumothorax has a high mor-bidity and seldom achieves its object, that is that itcan be abandoned when the underlying disease iscontrolled.
BIBLIOGRAPHYAYCOCK, J. B., BRANTIGAN, 0. C., and WELCH, W. (1940),
J. Thor. Surg., 9, 382.BAER, G. (1913), Munsch. Med. Wschr., 40, 1587.EDWARDS, F. R. (I949), Thorax, 4, 224.GRAF, W. (1936), Dtsch. Med. Wschr., 62, 632.GRENVILLE-MATHERS, R., and TRENCHARD, W. J. (1952),
Thorax, 7, I85.HOLST, J., SEMB, C., and FRIMANN DAHL, J. (I935), Acta
Chirg. Scand. Supplement 37.LAIRD, R. (I953), Lancet, 265, 319.LUCAS, B. G. B., and CLELAND, W. P. (I948), Thorax, I, 211.MORRISTON DAVIS, W., and TEMPLE, J., and STATHATOS,
C. (i95i), Ibid., 6, 209.MORRISTON DAVIS, W. (I933), 'Pulmonary Tuberculosis
Medical and Surgical Treatment,' London, Cassell.PRICE THOMAS, C., and CLELAND, W. P. (1942), Brit. J.
Tub., 36, I09.PRICE THOMAS, C. (1952), 'Modem Practice in Tuberculosis,'
Sellors, Livingstone & Butterworth.SAUERBRUCH, F. (I920), 'Die Chirurgie de Brustorgane.'SEMB, C. (I935), Acta Chirg. Scand. Supplement 37SELLORS, T. H. (I948), Thorax, 2, 2I6.SELLORS, T. H. (1953), personal communication.TUFFIER, T. (I895), Gaz Hosp. Paris, 68, 1320.WILSON, D. A. (1946), Surg. Clin. N. Amer., 26, Io6o.
Continued from bage 208-P. H. Scadding. M.D.. M.R.C.P.Acknowledgments
I am grateful to Mr. D. F. Kemp, of thephotographic department of the Institute of theDiseases of the Chest, for the X-ray photographsin Figs. I (a) to (d).
BIBLIOGRAPHYANDERSON, N. L., and WINN, W. D. (I945), Am. Rev. Tub.,
52, 380.BANYAI, A. L. (1946), 'Pneumoperitoneum Treatment,' Henry
Kimpton, London.
CROFTON, J. W. (I949), Thorax, 4, 96.EDGE, J. R. (I953), Brit. Y. Tuberc., 47, 202.FOSTER-CARTER, A' F., MYERS, M., GODDARD, D. L. H.,
YOUNG, F. H., and BENJAMIN, B. (1952), BromptonHospital Reorts, Vol. xxii.
KEERS R. Y. (948), Brit. 7. Tuberc., 42, 58.MORRIS, E., et al. (1952), J.A.M.A., 149, II20.RAFFERTY, T. N. (I947), 'Artificial Pneumothorax,' Grune &
Stratton, New York.RILANCE, A. B., and WARRING, F. C. (I944), Am. Rev. Tub.,
49, 353-SCADDING, F. H., NICHOLSON, H., and HOYLE, C. (x9S),
Quart. 7. Med., N.S. 2, 313.THOMAS, D. M. E. (I950), Personal communication.WRIGHT, G. W., et al. (I949), 7. Thor. Surg., I8, 372.VAJDA, L. (I,933), Ztschr. f. Tuberk., 67, 371.
