762

THE TREATMENT OF HAEMATEMESISG. N. CHANDLER, M.A., D.M., M.R.C.P.

Physician, Chapel Allerton Hospital, Leeds and Wakefield Area Hospitals:late Senior Registrar, Department of Gastroenterdlogy, Central Middlesex Hospital, London, N. W. io

Haematemesis is one of the commonest medicalemergencies that may threaten life. It is themost frequently encountered complication ofpeptic ulcer, occurring at some time in about 25%of cases,8 and peptic ulcer is easily the most im-portant cause of the admission of patients withsevere bleeding from the upper alimentary tract.In Avery Jones' series'2 of 2,526 admissions forthis emergency, 2,298 were due to proved orprobable peptic ulcers. Table i shows the dis-tribution of the various causes of bleeding inadmissions to the Central Middlesex Hospital forhaematemesis and/or melaena between 1940 andI957.

TABLE IPeptic ulcer group:

Chronic gastric ulcer .. ...... 423Duodenal ulcer ........ 844Post-operative group previous or partial

gastrectomy .. ........ I68'Acute lesion ' group ....... 720Hiatus hernia .. ..... . 47Unclassified (incompletely investigated) .. 96

2,298

Causes other than peptic ulceration:Carcinoma ventriculi .. .. .. .. 57Portal hypertension .. .. .. .. 75Other causes . .. .. .. 96

2,526

Diagnosis of the Cause of BleedingThough the probability is that the patient ad-

mitted after an attack of haematemesis and melaenais bleeding from a benign ulcerative process in theupper alimentary tract, there remains an importantgroup of less common causes including portalhypertension, gastric carcinoma and other tumoursof the gastro-intestinal tract, blood dyscrasias, andbleeding associated with the strain of vomiting(Mallory-Weiss syndrome) and general medicaldiseases such as periarteritis nodosa, uraemia andmalignant hypertension. Their recognition is not

only an interesting diagnostic challenge but alsoan important factor in subsequent management.The difficulty of treating bleeding peptic ulcer

largely concerns the detection of those patients inwhom the prognosis with medical measures mustbe considered poor and who will bleed to deathunless surgery is employed. The problem is notalways easily solved. Acute ulceration or erosionof the stomach or duodenum is a common lesionwhich usually responds well to medical treatment,but this diagnosis is not always obvious, beinglargely based on negative evidence and frequentlymade only in retrospect after a negative X-ray.The vessels by which an acute ulcer bleeds aremostly small and submucosal and operation shouldbe undertaken only if there has been severe re-current bleeding on two or three occasionsaccompanied by shock.

Extra-gastric sources of bleeding must also beexcluded before patients are submitted to surgery.The history is one of the most valuable aids inachieving a correct diagnosis; thus where chronicpeptic ulceration is responsible for the bleeding itwill be unusual to find such a patient denyingprevious periodic dyspepsia with food and alkalirelief of his pain. The shorter the history ofpreceding dyspepsia, the more likely is the lesionto be acute and, therefore, best treated medically;enquiry as to recent salicylate consumption maygive a clue to correct diagnosis, but unfortunatelythis is not always obvious and the history may bemisleading.There is thus much need for a method of

investigation that will yield diagnostic informationin the acute stage of illness, for if chronic pepticulceration can be diagnosed with confidence,surgical treatment is to be recommended inpatients over the age of 50 who bleed again afteradmission to hospital. At the Central MiddlesexHospital the following procedure of early com-bined investigation has been adopted. As soon asthe patient has been made comfortable afteradmission, a Ryle's tube is passed via the nose andabout 5 ml. of gastric contents are aspirated hourly

by copyright. on 29 June 2019 by guest. P

rotectedhttp://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.36.422.762 on 1 Decem

ber 1960. Dow

nloaded from

CHANDLER: The Treatment of Haematemesis

by syringe suction and pH estimated electro-metrically at the end of the intubation period,which finishes at 9 a.m. on the day after admission,when gastroscopy is performed. Later that morn-ing or in the early afternoon of the same day,barium meal examination is done on the ward usinga portable X-ray set and without manipulation orthe need to move the patient from his bed. Bycombining the information derived from study ofintragastric acidity,3' 4,5 from gastroscopy, andfrom early X-ray examination, it has proved pos-sible to achieve a correct diagnosis of the cause ofbleeding in 80% of patients admitted with haemor-rhage from peptic ulcer, usually within 24 or 36hours from admission.6 7 The reliability of theearly diagnostic procedures in forecasting correctdiagnosis was significantly greater than the clinicalimpression of the cause of bleeding in each of thethree groups of chronic gastric ulcer, chronicduodenal ulcer and acute ulcer.

General Considerations in TreatmentThough, to a large extent, the correct treatment

of haematemesis depends on the diagnosticprobabilities established by careful examinationand, where possible, confirmed by appropriateinvestigation, certain general measures apply toall cases of bleeding from the alimentary tract.There is no longer any controversy about theadvisability of liberal feeding. Meulengrachts'original diet4 has been modified in this country byWitts2 and it is this regime that is usually followed;most patients do well on a two-hourly soft dietwith fluids as desired between, but if nausea is aprominent complaint then two-hourly milk feedsare sufficient. There is no evidence that suchfeeding increases the liability to further haemor-rhage, the risk of which remains at about 25% ofadmissions for bleeding peptic ulcer.l2 Certainlyit is unphysiological to deprive an exsanguinatedpatient of food and fluid at such a time of specialneed and there is no doubt that modern treatmenthas eliminated the deaths which formerly occurredfrom therapeutic dehydration.

Strict confinement of the patient to bed isessential in the presence of severe or continuedbleeding, but there is no reason why the patientwhose course is uncomplicated should not beallowed up by the second or third hospital day.There is no evidence to suggest that recovery afterhaematemesis is hastened by confinement to bed,the dangers of which are enhanced in older patientswith associated degenerative diseases. There islittle doubt that, in the presence of satisfactoryclinical progress, the physician can best serve theinterests of his patients by allowing early ambula-tion, and Pollard and Summerskill18 have shownthat such a policy confers important benefits.

I An injection of sodium phenobarbitone, 200 mg.,may be required on admission to secure thepatients' mental relaxation; occasionally, mor-phine, o1 to 15 mg., may be needed for the anxietyand restlessness of shock but medication otherthan vitamin concentrates and antacids is bestavoided, though iron may be given by mouthfrom the start. Constipation is the rule afterhaematemesis and is usually well tolerated by thepatient. Purgatives should not be administeredbut there is no objection to the use of a glycerineor' Dulcolax' suppository. Aspirin should alwaysbe avoided and, instead, Panadol given forincidental pains or discomfort.

It is unusual for ulcer pain to persist afterhaemorrhage has occurred and continued severepain suggests the possibility of concomitant per-foration (a particularly lethal combination) or ofan independent cause such as cardiac infarction.A continuous intra-gastric milk drip, giving 6pints in 24 hours through a small nasal tube, is anextremely valuable treatment if the symptoms ofactive peptic ulceration recur.Though brisk continued bleeding from peptic

ulcer is an indication for surgical treatment, thereare a few patients in whom operation can reason-ably be considered as out of the question becauseof severe associated disease or the features ofoverwhelming senility. Nevertheless, howeverunpromising the prognosis may appear, neitherhope nor treatment should be abandoned. Dis-tension of the stomach with blood clot may be animportant cause preventing the atonic stomach of ashocked patient from arresting the bleeding bycontraction. Emptying the stomach with aSenoran's evacuator followed by lavage with ice-cold-water may help to stop bleeding and the useof a topical haemostatic such as thrombin afterwashing the stomach out with I:I,000 adrenalinhas been advocated in the treatment of bleedingacute peptic ulcer. In such cases the stomachshould first be emptied through a large-bore tubeand then lavaged with : ,000o adrenalin; finally,thrombin in a suitably viscous vehicle such asmethyl cellulose is instilled into the stomach.

Blood TransfusionAn hourly record should be kept from admission

of pulse and blood pressure, and blood immediatelytaken for haemoglobin or haematocrit and into aplain tube for grouping and cross-matching fortransfusion. There can be no hard-and-fast rulesas to when blood transfusion is required. Whilstblood volume determinations would enable theclinician accurately to assess the amount of bloodlost, such techniques are, as yet, unsuitable forroutine use and reasonable working rules for theemployment of blood transfusion are a pulse rate

December i96o 763by copyright.

on 29 June 2019 by guest. Protected

http://pmj.bm

j.com/

Postgrad M

ed J: first published as 10.1136/pgmj.36.422.762 on 1 D

ecember 1960. D

ownloaded from

POSTGRADUATE MEDICAL JOURNAL

of 20o or more, a systolic blood pressure ofo00 mm. Hg. or less, or if the haemoglobin isbelow 50%. Tibbs24 found that the diastolic bloodpressure reflected the diminution in blood volumemore accurately than the systolic pressure, exceptin hypertensive patients who maintain theirdiastolic pressure better than normals in thepresence of bleeding. A diastolic pressure of60 to 65 mm. Hg. indicated moderate to severeblood loss. Transfused blood does not act as ahaemostatic but there is no evidence to suggestthat it provokes recurrence of bleeding. It is bestregarded as a remedy for haemorrhagic shock andas giving the patient a further margin of safetyshould he bleed again. Prompt transfusion isparticularly important in the elderly to prevent theirreversible cerebral damage that may complicateprolonged shock. If severe anaemia is allowed todevelop a compensatory increase in venous pres-sure leads to greater cardiac filling and improvedcardiac output. This hyperkinetic phase presentswith a full bounding pulse, raised pulse pressureand jugular venous distension. The venous pres-sure also rises during the initial stages of trans-fusion, and though it falls again as the arterialpressure rises, overtransfusion in the hyperkineticphase carries the risk of precipitating heart failure.

Usually gastro-duodenal bleeding stops within12 or 24 hours of admission, most often per-manently, and during this time I,500 to 3,000 ml.of blood may be given by slow drip transfusion(40 drops per minute). Occasionally there is acontinuous slow loss of blood requiring intermit-tent transfusion and this seems to happen mostcommonly in patients bleeding from acute lesions;the vessels by which an acute ulcer bleeds aremostly small and submucosal, and such bleedinglacks the dramatic quality of arterial haemorrhagefrom the base of a chronic ulcer. Exceptionally,chronic ulcers bleed so severely that the only hopefor the patient lies in intra-arterial transfusion andimmediate operation.

Transfusions of 500 ml. are unnecessary andwasteful. Large volumes of blood given rapidlycarry risks of citrate intoxication and hyper-kalaemia. The former may lead to defectiveclotting and myocardial failure and it is wise togive io ml. of io% calcium gluconate intra-venously after every fourth bottle of blood.Potassium intoxication carries the very real threatof cardiac arrest, particularly if bank blood nearthe limit of its expiry is used, in which serumpotassium levels may reach 25 mEq/l.13 Warmingthe blood before use will encourage the return ofpotassium to the cells and the danger can befurther diminished by the infusion of dextrosesolutions. However, when such massive trans-fusions are in question, the need for surgical

arrest of haemorrhage should be urgently re-viewed.

Surgery of Bleeding Peptic UlcerThough the majority of patients recover under

medical management, there is undoubtedly asignificant proportion who will bleed to deathunless surgery is employed. If chronic pepticulceration can be diagnosed with reasonableconfidence, operative treatment is to be recom-mended for those older patients in whom bleedingis severe, or recurs after admission to hospital. Asearly as 1918, Finsterer9 proposed the applicationof radical surgical measures, the operation beingperformed within the first 24 to 48 hours from theonset of haemorrhage; his reported operativemortality for early cases was 5%. Gordon-Taylor10commented: ' Finsterer's first 48 hours is still theoptimum period for surgical attack in haematemesisand the golden age of gastric surgery will havebeen attained only when all cases of haemorrhagefrom chronic ulcer come to operation within thatspace of time '. And Tanner23 found that the bestresults were obtained by early and frequentsurgery. While the mortality with medicalmanagement can be as low as the 2.5% achievedby Meulengracht,15 it is unlikely that a universalsurgical approach will provide either comparablesuccess or achieve wide acceptance. In general,surgical intervention has been limited to selectedpatients2 16, 17 the best results being achieved byprompt surgery in patients with proved or probablepeptic ulcers. Avery Jones," in a study of 400consecutive admissions for haematemesis, foundparticular indications for emergency gastrectomyin patients over 50 years with good clinical evidenceof chronic ulcer who had brisk recurrent haemor-rhage after admission. Probably the most im-portant factor influencing mortality in haema-temesis is the age of the patient; above 60 yearsthere is a steep increase in the proportion of deathsfrom bleeding peptic ulcer and it is in these casesthat the best surgical contribution to the reductionof mortality can be made.With awareness of the factors influencing

mortality in bleeding peptic ulcer, a policy ofselective surgical intervention was introduced atthe Central Middlesex Hospital in I946. Surgerywas readily considered in patients known to havea chronic ulcer if they bled again after admission.No patient was denied operation in the face ofsevere recurrent bleeding, but a much strongerbias was exerted in favour of operation over theage of 40. Surgical treatment was undertaken inpatients with a presumed or proven acute ulcer ifthere was severe recurrent bleeding on two or threeoccasions accompanied by shock. Although bleed-ing acute ulcer has a very low mortality in younger

764 December I 960by copyright.

on 29 June 2019 by guest. Protected

http://pmj.bm

j.com/

Postgrad M

ed J: first published as 10.1136/pgmj.36.422.762 on 1 D

ecember 1960. D

ownloaded from

December 1960 CHANDLER: The Treatment of Haematemesis 765

TABLE 2

HAEMATEMESIS AND MELAENAANNUAL MORTALITY-PEPTIC ULCER GROUP

Year No. Deaths Mortality

941 52 5.7% Medical1942 71 2 policyI943 92 6.6%944 io6 91945 124 8

'947 9'I948 13I 7.4% Selective sur-1948 113 6 gical policy1949 136 9 9.4%1950 119 151951 203 131952 I88 I6 7.4/1953 I79 16

1955 I86 71956 153 7 4.1%1957 195 8J

2,302 1-57

people, the mortality is appreciable over the ageof 60.12As Tanner22 has stressed, the criterion for

success or failure of a surgical policy is its influenceon overall mortality and not the mortality forsurgical intervention as such, which depends onthe type of risk accepted. The result of theadoption of selective surgical treatment at theCentral Middlesex Hospital is shown in Table 2which indicates that the overall mortality frombleeding peptic ulcer has been reduced to 4%.

Further analysis of the figures showed thatpatients with recurrent bleeding at any age fromgastric ulcers, and those bleeding from duodenalulcers over the age of 6o have particularly benefitedfrom a surgical policy. Chronic gastric ulcer is aparticularly strong indication for surgery; bleedingfrom this source carries a high mortality undermedical treatment and its operative arrest bygastrectomy is generally easier than is the surgeryof bleeding chronic duodenal ulcer. Though thesefigures cover only those patients with proved orprobable peptic ulcer, there are many other pos-sible causes of severe gastro-intestinal haemor-rhage which may require surgical intervention,such as tumours of the stomach, vascular lesionsof the intestine, diverticular disease, and portalhypertension.The success of any surgical approach to the

treatment of haematemesis requires close co-operation between physicians and surgeons if, asseems proper, patients continue to be admitted tomedical wards. In every case all evidence ofchronic peptic ulcer from previous investigationmust be considered in conjunction with the history

before admission in order to assist the final choicebetween operative and conservative treatment.The adoption of a policy whereby early diagnosisof the cause of bleeding can be achieved",7 greatlyfacilitates correct appraisal.

If operation be decided upon, a little time may beallowed to improve the patients' condition by bloodtransfusion unless bleeding be so profuse as topermit no delay. The surgical treatment of bleed-ing peptic ulcer requires an experienced gastricsurgeon and preferably a senior anaesthetist.There is a risk of aspiration into the lungs and it isimportant that a cuffed endotracheal tube beinserted quickly. It is equally important for thesurgeon to remove all blood clot from the stomachat operation to minimize the danger of post-operative vomiting and inhalation.

Partial gastrectomy is the operation of choice inall patients; local suture of bleeding points, orlocal excision of a gastric ulcer have little to recom-mend them. Any lesser operation than gastricresection is usually futile. There is a difficultproblem facing the surgeon who is forced tooperate in the face of massive recurrent haemor-rhage when no lesion can be identified. Theabdomen should never be closed if the surgeonis satisfied that the blood is coming from theupper reaches of the gut and not from lower down;either a wide gastrotomy for inspection and pal-pation of the gastric mucosa should be made-aprolonged and sometimes uncertain procedure-or a 'blind' gastrectomy performed. In thelatter case the ulcer responsible for bleeding willbe found in almost every patient on opening theresected specimen.Tracheotomy and Gastrectomy

Severe chronic bronchitis and emphysema is aconstantly recurring clinical accompaniment ofhaematemesis. Such cases can present formidableoperative risks and so bias the physician againstoperation that medical measures are persistedwith despite every evidence that only surgicaltreatment can stop the bleeding. It is in thesecases of severely impaired respiratory functionthat tracheotomy, done at the time of gastrectomy,can transform the post-operative outlook by mak-ing efficient bronchial drainage possible withintermittent suction and by improving ventilation.Experience at the Central Middlesex Hospital in asmall number of cases has been encouraging, andin most instances the tracheotomy tube can beremoved in 7 to io days. An alternative approach,successfully applied by Schooling and Simon,20was by use of artificial respiration with a BeaverMark 2 respirator connected to a short cuffedendotracheal tube inserted via a tracheotomythrough which bronchial toilet was performed with

by copyright. on 29 June 2019 by guest. P

rotectedhttp://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.36.422.762 on 1 Decem

ber 1960. Dow

nloaded from

POSTGRADUATE MEDICAL JOURNAL

intermittent suction. This method has the ad-vantage over tracheotomy alone of ensuringadequate respiration in patients whose ventilatingfunction depends almost entirely on diaphragmaticmovement, and who pass into severe respiratorydistress after an upper abdominal operation.Treatment of Bleeding Oesophageal VaricesThe development and subsequent rupture of

oesophageal varices constitute the most importantcomplication of portal hypertension. In Britainemergency admission to hospital because of bleed-ing from this cause is relatively infrequent (3% ofall admissions to the Central Middlesex Hospitalfor haematemesis and melaena) compared withAmerican experience where as many as 12.5% ofadmissions for upper gastro-intestinal haemor-rhage have been ascribed to portal hypertension.19

Bleeding from oesophageal varices is usuallysevere and exsanguinating, most often originatingin the region of the oesophago-gastric junction. Inpatients whose portal hypertension depends onextra-hepatic portal vein obstruction the prognosisis good, and once bleeding has stopped there isusually a rapid return of good health. In patientswith cirrhosis, however, haemorrhage is notoriouslydifficult to control and is only too often followed bydeepening jaundice, ascites and coma. There is,therefore, every reason to arrest haemorrhage incirrhosis with the least delay.Temporary control of haemorrhage from oeso-

phageal varices can be obtained in most cases bymeans of compression with the triple-lumenSengstaken-Blakemore tube. This is inserted untilthe gastric balloon is well below the cardia andinflated with I00 ml. of water. The inclusion of20 ml. ' Hypaque' contrast medium enables itsposition in the stomach to be easily located byportable plain X-ray film of the abdomen andgentle traction on the tube is sufficient to anchorthe balloon at the cardia, this position beingmaintained by strapping to the cheek with adhesiveplaster. Traction on the tube using a pulley systemand weights may be dangerous and is unnecessary.The oesophageal balloon is next inflated to apressure of 20 to 30 mm. of mercury. The gastricballoon compresses the upper part of the stomachagainst the diaphragm and controls bleeding fromthe submucosal plexus of veins in the region of thecardia whilst the oesophageal balloon helps toanchor the tube in addition to exerting directpressure on oesophageal varices. The third lumenof the tube should be used for feeding the patientwith 20% glucose solution, for instilling neomycin,potassium supplements and purgatives and fordiagnostic aspiration. If bleeding continues, asjudged by blood staining of the aspirated gastriccontents, it is reasonable to assume that the

haemorrhage is originating elsewhere, often froman associated gastric or duodenal ulcer. This maysometimes be confirmed by the introduction of aradio-opaque medium such as ' Gastrografin ' intothe stomach via the tube, using the techniquedescribed by Chandler et al.7

It is always difficult to decide on the appropriatetime for removal of the tube. The procedure isnot without risk, and the dangers of asphyxiafrom the inflated balloon slipping up into thepharynx and of gastric and oesophageal ulcerationincrease the longer the tube is left in position.Generally the oesophageal balloon is deflatedafter 24 hours and it is advisable that decompres-sion be carried out in the morning so that if thereis recurrent haemorrhage this can be dealt with asa day-time emergency and not in the early hours.Before withdrawal it is wise policy to leave theapparatus lying in situ before aspirating the gastricballoon and removing the tube, to ensure thatbleeding has stopped.There is perhaps no other emergency in which

general medical treatment, in addition to themeasures adopted to control bleeding, are of suchimportance. Adequate blood transfusion is essen-tial and oxygen may be required to combatcerebral anoxia. Hypoprothrombinaemia may bea factor disturbing the clotting mechanism in thesepatients and vitamin K should be administeredparenterally. The prevention of coma is of firstimportance and this risk can be minimized byreducing intestinal bacterial activity with neomycin2 g. four times a day. The administration of anenema will help to remove as much as possible ofblood from the bowel. Both diuretics1 and pituit-rin21 have been shown to be effective in loweringportal pressure and both are worthy of trial.The patient who continues to bleed despite all

medical measures presents a formidable problem,for there is no reliable emergency surgical measurefor dealing with life-threatening haemorrhage.Indeed, in many patients the bleeding is only oneof many abnormalities incidental to liver failureand a fatal termination. The construction of aporto-caval shunt is a long procedure which theseverely-ill cirrhotic patient is poorly fitted towithstand. Probably some lesser procedure, suchas ligation of the varices, or Tanner's operationsin which the left gastric vessels are ligated anddivided, the upper stomach transected and theends anastomosed again, is the best method ofachieving a more than temporary control of thesituation, but the mortality is still high. Shouldthe patient's condition improve sufficiently, con-sideration can then be given at a later date to avenous shunt operation but, despite early en-thusiasm for these procedures, it seems likely thatthe prognosis of the bleeding cirrhotic patient is

766 December I960by copyright.

on 29 June 2019 by guest. Protected

http://pmj.bm

j.com/

Postgrad M

ed J: first published as 10.1136/pgmj.36.422.762 on 1 D

ecember 1960. D

ownloaded from

December 1960 CHANDLER: The Treatment of Haematemesis 767determined by the disease in the liver rather thanby the pressure in the portal vein.

REFERENCESI. ATKINSON, M. (1959), Lancet, ii, 819.2. BOHN, G. (I949), Brit. med. J., ii, 630.3. CHANDLER, G. N., and WATKINSON, G. (1953), Lancet,

ii, 1170.4. CHANDLER, G. N., and WATKINSON, G. (x958), Quart.

J. Med., 27, 564.5. CHANDLER, G. N., and WATKINSON, G. (1959), Ibid.,

28, 371.6. CHANDLER, G. N., CAMERON, A. D., NUNN, A. H., and

STREET, D. F. (I96oa), Gut, I, 6.7. CHANDLER, G. N., CAMERON, A. D., NUNN, A. H., and

STREET, D. F. (I96ob), Lancet, ii, 507.8. CROHN, B. B. (I950), Univ. West. Ont. med. J., 20, 98.9. FINSTERER, H. (1939), Surg. Gynec. Obstet., 63, 291.

o1. GORDON-TAYLOR, G. (1935), Lancet, ii, 81i.i . JONES, F. AVERY (1947), Brit. med. y., ii, 441.

12. JONES, F. AVERY, and GUMMER, J. W. P. (I960),' ClinicalGastroenterology '. Oxford: Blackwells.

13. LEVEEN, H. H., PASTERNACK, H. S., LUSTRIN, I.,SHAPIRO, R. B., BECKER, E., and HELFT, A. E. (1960),J. Amer. med. Ass., 173, 770.

14. MEULENGRACHT, E. (1935), Lancet, ii, 1220.IS. MEULENGRACHT, E. (I947), Atch. int. Med., 80, 697.I6. PARSONS, K. O., and ALDRIDGE, L. W. (i95i), Brit. J.Surg., 38, 370.17. PEDERSEN, J. (I95s), Lancet, i, 1292.I8. POLLARD, A., and SUMMERSKILL, W. H. J. (x960),Brit. med. J., i, 171.I9. SCHIFF, L., and SHAPIRO, N. (1951), 'Peptic Ulcer'.

Philadelphia: W. B. Saunders Co.20. SCHOOLING, I. B., and SIMON, L. (1958), Brit. J. Surg.,

46, 86.z2. SCHWARTZ, S. I., BALES, H. W., EMERSON, G. L., and

MAHONEY. E. B. (I959), Surgery, 45, 72.22. TANNER, N. C. (1950), Proc. roy. Soc. Med., 43, I47.23. TANNER, N. C. (1954), Postgrad. med. J., 30, 577.24. TIBBS, D. J. (1956), Lancet, ii, 266.25. WITTS, L. J. (1937), Brit. med. 7., i, 847.

HAEMATOLOGY(Postgraduate Medical Journal, September 1959)

Price 6s. 6d. post free

THE DIFFERENTIAL DIAGNOSIS OF THE DGNOSIS AND MAAGEMEN OFHAEIMOYTICV AEMIAFA5 1THE DIAGNOSIS AND MANAGEMENT OFHAEMOLYTIC HAEMOPHILIA AND CHRISTMASGeorge Discombe, M.D., B.Sc. DISEASETHE RELATIONSHIP BETWEEN GENETIC, M. C. G. Israels, M.Sc., M.D., F.R.C.P.,NUCLEAR AND SOCIAL SEX F. Nour-Eldin, Ph.D., M.B., L.S.A., and

William M. Davidson, M.D., and Shirley John F. Wilkinson, M.Sc., Ph.D., M.D.,Winn, B.Sc. F.R.C.P., F.R.I.C.

CELL COUNTSJ. W. Stewart, M.B., B.S., and P. J. Crosland- HAEMORRHAGIC STATESTaylor, M.B., B.Chir. E. K. Blackburn, M.D., F.R.F.P.S.

THE ASSESSMENT OF PLATELETFUNCTION ATYPICAL PERNICIOUS ANAEMIAA. A. Sharp, M.D., B.Sc. Allan Jacobs, M.D.

Published byTHE FELLOWSHIP OF POSTGRADUATE MEDICINE

9, Great James Street, London, W.C.1

by copyright. on 29 June 2019 by guest. P

rotectedhttp://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.36.422.762 on 1 Decem

ber 1960. Dow

nloaded from