article the - postgraduate medical journal142 a.e. gimson&d. westaby rate in controlling...

Postgrad Med J (1991) 67, 140- 146 i) The Fellowship of Postgraduate Medicine, 1991

Review Article

The management of an episode of variceal bleeding

A.E. Gimson and D. Westaby

Liver Unit, King's College School ofMedicine and Dentistry, Denmark Hill, London SE5 8RX, UK

Introduction

Variceal bleeding is the most serious consequenceof portal hypertension and may be anticipated atsometime during the lifetime in 30% of patientswith an underlying cirrhosis.' The admission mor-tality for an episode ofvariceal bleeding is criticallydependent upon the severity of the underlying liverdisease but may be as high as 50% for the firsthaemorrhage.' In the absence of specific therapy toprevent it, recurrence of bleeding will occur in20-60% of patients during the period of admis-sion, the considerable variation, in part, reflectingdifferent definitions of this event.2-4 The severity ofthe liver disease is the most important factorpredicting rebleeding.2'3 The specific therapy for anepisode of variceal bleeding is quite distinct fromthat required for other causes of gastrointestinalbleeding and as such this places considerableemphasis upon the need for early endoscopicdiagnosis following initial resuscitation. At thetime of diagnostic endoscopy (carried out within 4hours ofadmission) fewer than 50% ofpatients willbe actively bleeding, the remainder having stoppedspontaneously.2'5 This distinction is important inthat the measures directed towards the control ofactive bleeding may not be the same as those usedto prevent early recurrence (in those who havestopped spontaneously - see below).

Resuscitation

Standard resuscitative measures for volume deple-tion are applicable in patients with variceal haem-orrhage and will not be covered in detail. However,there are a number of considerations, specific topatients with cirrhosis, which deserve mention. Thenormal response to hypovolaemia includes bothvenous and arteriolar constriction, mediated init-ially by the low pressure baro-receptor reflex,thereby maintaining venous return and redistribu-tion ofblood flow to vital organs. In the presence of

cirrhosis, particularly severe disease, the baro-receptor reflux may be diminished by an attenuatedresponse of the alpha-receptors to noradrenaline.6Furthermore, the ability to mobilize the so-called'unstressed volume' of blood from the hepaticsinusoids to maintain venous return may also berestricted in the cirrhotic liver.' Both these factorsmay lead to a greater fall in arterial pressure thanwould be anticipated from the observed blood lossand emphasizes the need for prompt resuscitation.The degree to which volume restitution is under-

taken has been influenced by the observation thathyper-expansion of the circulation may precipitatevariceal bleeding.8'9 Two recent studies have assess-ed the effect of volume depletion and repletion onportal and systemic haemodynamics in rat modelsof portal hypertension.'0"' Venesection (20% ofblood volume) produced a significant reduction inportal pressure; however, this then rose abovebaseline values when the exact volume ofblood wasreplaced.'0 The degree of 'overshoot' in portalpressure appeared to correlate with the extent ofthe collateral circulation and was secondary to arise in portal collateral resistance." Whether asimilar elevation of portal pressure occurs in manand to what extent this might represent a clinicalproblem in the form of protracted or recurrentvariceal bleeding is unknown. The need to protectrenal function does not permit inadequate volumerestitution although overfilling should be avoidedby titration against central venous pressure. Cau-tion should be taken when interpreting the centralvenous pressure measurements in patients withtense ascites: compression of the right atrium maylead to an overestimate of the value. 2 The measure-ment of the pulmonary capillary wedged pressureusing a pulmonary flotation catheter will provide amore accurate assessment in such cases.

Therapy of active variceal bleeding

The measures used to manage an episode ofvariceal bleeding may be conveniently categorized

Correspondence: D. Westaby, M.A., M.R.C.P.Received: 13 August 1990

copyright. on M

ay 11, 2021 by guest. Protected by

http://pmj.bm

j.com/

Postgrad M

ed J: first published as 10.1136/pgmj.67.784.140 on 1 F

ebruary 1991. Dow

nloaded from

http://pmj.bmj.com/

MANAGEMENT OF VARICEAL BLEEDING 141

into those that have an effect only for the periodthey are applied, those which may prevent early(hours to days) rebleeding and those which mayreduce the risk of long-term (weeks to months)recurrent haemorrhage (Table I).

This distinction has important clinical implica-tions in that therapy of only short-term benefitshould be followed by a more definitive procedureto prevent the high incidence of rebleeding.

Table I The management of variceal bleeding

1) Measures effective only during applicationVasocontrictor agents

Drugs increasing lower oesophageal sphincter pressureBalloon tamponade

2) Measures with additional effect on early rebleedingInjection sclerotherapy

Banding ligation(Qesophageal transection without devascularization)

3) Measures with additional long-term effect on rebleedingPortal systemic shunt surgery

Ablative surgery

1. Measures effective only during the period applied

Vasoconstrictor therapy

(a) Vasopressin

Despite almost 40 years of continuous use onlyfour controlled trials of vasopressin, as comparedto placebo or 'conventional' therapy, have beenreported.'3-'6 In the first three of these studies asignificant benefit was observed for vasopressinwith haemostasis being obtained in 44-71% ofcases.'3 15 However, in the most recent trial vaso-pressin failed to show significant benefit over aplacebo infusion (59% vs 45% CI-9- + 39% re-spectively).'6 Similar low success rates (9-58%) forvasopressin have been observed in a number ofstudies in which this has been compared to otherpharmacological agents.'7-'9 Attempts to improvethe efficacy of vasopressin by direct administrationinto the superior mesenteric artery failed to showbenefit when compared to a continuous intra-venous infusion.'4"5Two major limitations of vasopressin therapy

can be recognized from the available studies: firstly,the frequency of complications associated with itsgeneralized arterial vasoconstrictor effect whichhave required cessation of therapy in 20% of casesand death in 3% (the majority from myocardialinfarction or left ventricular failure); secondly,even in those patients who tolerate the drug,

control of bleeding has been achieved in only50-65% of cases.20

(b) Vasopressin combined with a vasodilator

The concept of combining vasopressin and avasodilator was conceived with the aim ofreducingthe cardiovascular complications associated-withthe former and to thereby enable a greater propor-tioU of patients to be exposed to the potentialbenefits of the therapy. Both venous and arterialvasodilators have been investigated and shown toreverse the systemic haemodynamic effects ofvaso-priessin whilst the fall in portal pressure wasmaintained or enhanced.21-2' These changes de-pend upon the vasodilatation of selective areas ofthe systemic circulation without reversing the vaso-pressin effect on the splanchnic arteries. The pre-dominant venodilators, such as nitroglycerin, mayalso act by reducing pre-load and thereby protect-ing the left ventricle in the presence of vasopressininduced coronary vasoconstriction.

Three controlled trials have been reported com-paring the combination ofvasopressin and nitrogly-cerin with vasopressin alone for the control ofactivevariceal bleeding.24-26 The route by which the nitro-glycerin was administered differed in each study andincluded the intravenous,24 sublingual25 and trans-dermal approach.26 Each of these three trials pro-vided evidence of benefit for the combined regimewith respect to either enhanced control ofbleeding26or reduced complications25 or both.'4 The observedbenefits did not result in an improvement in admis-sion mortality. This may in part be explained by thefailure of the combined regime to improve thecontrol of the most severe episodes of bleeding.24Furthermore, GradeC patients with the most severeliver disease (Chill's Grade C) account for themajority ofdeaths and in the presence ofsevere liverdisease it may be the fact of bleeding rather than theextent that is the major determinant of outcome.

Despite the failure of the combined regime toinfluence mortality as compared to vasopressinalone, the enhanced control of bleeding (withevidence of reduced transfusion requirements)2426and fewer complications may be anticipated todiminish the resources required to manage activevariceal bleeding.

(c) Triglycyl-lysine vasopressin (glypressin)

Glypressin, the triglycyl-lysine analogue of vaso-pressin, was introduced as a successor to vasopres-sin for the management of variceal haemorrhage.2'The ability to administer the drug as a peripheralintravenous bolous at 6-hourly intervals offered amajor advantage as compared to the continuouscentral venous infusion required with vasopressin.Early uncontrolled studies suggested a high success

copyright. on M


http://pmj.bm

j.com/

Postgrad M


ebruary 1991. Dow

nloaded from

http://pmj.bmj.com/

142 A.E. GIMSON & D. WESTABY

rate in controlling bleeding and with few sideeffects.28 A controlled study comparing glypressinwith vasopressin confirmed enhanced control ofbleeding with the former (70 vs 9%, P<0.02)although the number of patients included in thestudy was extremely small (20) and the efficacy ofvasopressin considerably less than would be pre-dicted.'7 The same authors then went on to carryout a placebo based controlled trial and in aninconclusive report showed benefit for glypressin(60 vs 37%), this difference failing to reachsignificance (P< 0.05). ' Evidence in support oftheefficacy ofglypressin comes from the largest report-ed placebo based controlled trial,29 although theroutine use of balloon tamponade and injectionsclerotherapy in the patients included precludes aconfident assessment of benefit.A recent study has reported the effects of

combining glypressin with a venodilator (nitro-glycerin) and confirmed a beneficial effect upon thesystemic haemodynamic changes observed with theformer;30 the addition of nitroglycerin had no effectupon the portal pressure reduction produced byglypressin. The efficacy of glypressin and nitro-glycerin for the control of variceal bleeding hasbeen assessed in a single controlled trial in whichthis combination was shown to be comparable toballoon tamponade (78% and 79% respectively 12hours after institution of therapy).3' The additionof nitroglycerin to glypressin therapy was welltolerated and no serious cardiovascular complica-tions were observed.

(d) Somatostatin

Somatostatin has been shown to reduce portalinflow (mediated by splanchnic artery vasocon-striction) and hence portal pressure without induc-ing adverse systemic haemodynamic effects.32 Thefirst two controlled trials compared somatostatinto vasopressin as a single agent;'833 both reportedfewer side effects with somatostatin although inonly one study was there enhanced control ofbleeding.33 Two placebo based controlled trialshave recently been reported with markedly con-trasting results.3435 In the largest of all the reportedstudies (120 bleeding episodes) haemostasis wasobtained more frequently with somatostatin thanplacebo (64 and 41% respectively; P< 0.05).34 Thesecond placebo based study incorporated a similartrial design and included 84 patients, all of whomhad evidence of recent major haemorrhage.35 How-ever, the control of bleeding at 30 hours afterrandomization was less in the somatostatin groupthan those receiving placebo (65 and 83% respec-tively; P = 0.06). A most unexpected feature ofthese results is the extremely high spontaneouscessation of bleeding in the placebo group whichcould not have been predicted from a review of

established literature. Until further trials have beenreported it is difficult to justify the costs incurred byusing somatostatin.

Drugs affecting the lower oesophageal sphincterpressure

Drugs which increase lower oesophageal sphincterpressure such as metoclopramide, domperidoneand pentagastrin have been shown to reduceazygos blood flow36 and to lower varix pressure.37There is preliminary evidence to suggest that thesedrugs might provide temporary control of varicealbleeding and thereby facilitate measures such asinjection sclerotherapy.3"

Balloon tamponade

Direct compression ofthe bleeding point by the useofballoon tamponade has been in use for almost 40years.38 Available reports suggest control ofvariceal bleeding may be anticipated in 80-90% ofcases with an approximate 3% mortality directlyassociated with the tube.20 However, such resultsare the product of centres experienced in the use ofthe technique. A recent report suggests that theefficacy ofballoon tamponade may be considerablyless and the complication rate higher when appro-priate expertise is lacking.39 The optimum design ofthe tube has been the subject of considerabledebate. The use ofa 4 lumen tube (incorporating anoesophageal aspirate port) has been shown toreduce the risks of aspiration pneumonia as com-pared to the earlier 3 lumen Sengstaken-BlakemoreTube.' A second important consideration is theneed to use the oesophageal as well as the gastricballoon. Inflation of the oesophageal balloon hasbeen shown to have minor benefits with respect tohaemostasis but with an increased risk of aspira-tion pneumonia.4" 42 There is evidence to suggestthat the single gastric balloon tube (Linton-Nach-las) provides better control of gastric varicealbleeding.4' However, any observed benefit is mostlikely to reflect the volume of the gastric balloon(the Linton-Nachlas Balloon can be inflated to600 ml); any tube in which inflation of the gastricballoon can attain approximately 200 ml shouldprovide similar control. The critical factor forcontrol of both oesophageal and gastric varicealbleeding is the close aposition ofthe gastric balloonto the gastro-oesophageal junction.39

Balloon tamponade inevitably leads to mucosaldamage particularly at the gastro-oesophagealjunction and this becomes more marked withprolonged application. These lesions may become a,subsequent cause of bleeding when the tube isremoved. A period not exceeding 12 hours willminimize the risks of serious damage.

copyright. on M


http://pmj.bm

j.com/

Postgrad M


ebruary 1991. Dow

nloaded from

http://pmj.bmj.com/


2. Measures with additional effect upon earlyrebleeding

Injection sclerotherapy

3. Measures with additional influence upon laterebleeding

Portal systemic shunt surgery

Support for the use of injection sclerotherapy foran episode of variceal bleeding first came from twolarge uncontrolled studies in which control ofbleeding for the period of admission was in excessof 90%.4"3 The majority of the patients includedhad either stopped bleeding spontaneously or hadinitially been managed by balloon tamponade andas such no conclusions may be drawn with respectto sclerotherapy as single treatment for activehaemorrhage. A number of controlled trials havenow been reported comparing sclerotherapy as asingle therapy with balloon tamponade and/orvasoconstrictor therapy for the period of admis-sion and provide strong evidence to support theefficacy of the former.45-49 However, much of thebenefit associated with sclerotherapy may derivefrom the prevention of early rebleeding, an attri-bute which is not inherent to vasoconstrictortherapy or balloon tamponade. A recent study hasspecifically investigated the efficacy ofsclerotherapy for active variceal bleeding in com-parison to a regime of vasopressin and nitro-glycerin.' Control of bleeding was assessed over 12hours, after which both groups were treated bylong-term sclerotherapy (thus removing the biasthat may occur if this was restricted to thesclerotherapy group alone). Haemostasis at 12hours was obtained in a significantly higher pro-portion ofpatients treated by sclerotherapy (88 and65% respectively, P< 0.05) with lower transfusionrequirements. Overall frequency of rebleeding andadmission mortality were not different, almostcertainly reflecting the use of long-termsclerotherapy in both groups.

In 15-20% ofcases variceal bleeding arises fromgastric varices.5 Those situated on the lesser curveor within a hiatus hernia may be managed bysclerotherapy with similar efficacy to that attainedfor oesophageal varices.5152 The results of sclero-therapy for bleeding from fundal varices is poorand there should be an early recourse to surgery.53In the presence of a bleed-free interval fundalvarices may be successfully treated by the intra-varix injection of a tissue adhesive.53

Endoscopic banding ligation ofoesophageal varices

The recent introduction of an endoscopic techni-que enabling the direct placement of pre-stressedrubber bands upon the varices represents an alter-native to injection sclerotherapy as a local measurefor the management of variceal bleeding. Initialresults of this technique appear to justify furtherevaluation in controlled trials.'

Support for shunt surgery as an immediatemeasure (within 8 hours of admission) for varicealhaemorrhage comes from a large uncontrolledseries in which 180 unselected patients (althoughonly 15% Grade C) received a portal caval anasto-mosis.55 Control ofbleeding was attained in 98% ofcases but at the expense of a 42% admissionmortality. Despite this high admission mortality30% of the patients were alive at 12 years. Fromthese data it appears that the use ofshunt surgery asa first line measure selects out patients who, withprompt control of bleeding and absence of recur-rence haemorrhage, have a good prognosis.

In a recently reported controlled trial an 'emer-gency' portal caval shunt was compared to injec-tion sclerotherapy (immediate and long-term) inpatients with Grade C (Cello Modification) liverdisease.56 During the period of admission 50% ofthe sclerotherapy group rebled compared to 19%of those treated by shunt surgery (P< 0.05).Admission mortality was similar (50 and 56%sclerotherapy and shunt surgery respectively).Contrary to expectations the occurrence of ence-phalopathy was not different for the two groups.

Despite this evidence in support of immediateshunt surgery, the high admission mortality hasconsistently deterred more widespread adoption infavour of its use in patients who have failed othertherapies. The value of 'rescue' shunt surgery hasreceived support from the results of a controlledtrial in which a selective distal spleno-renal shuntwas carried out for rebleeding in patients treatedinitially by sclerotherapy.57A further factor which has an important bearing

upon surgical intervention is the prospect of livertransplantation. Although there are no absolutecontra-indications to the different shunt opera-tions, any previous upper abdominal surgery in-creases the risk of transplantation.58 The centralspleno-renal and meso-caval shunts, which areconstructed at a distance from the liver hilum, mayhave advantages over the portal caval shunt.

Ablative surgery

The adaptation of the surgical stapling gun fortransection of the oesophagus has led to a resur-gence of interest in ablative therapy for the man-agement ofactive variceal bleeding. This procedurehas been used alone or in conjunction with devas-cularization. The extent of the devascularizationhas important implications as to the suitability of aprocedure for 'emergency' use.

Three controlled trials have compared oeso-

copyright. on M


http://pmj.bm

j.com/

Postgrad M


ebruary 1991. Dow

nloaded from

http://pmj.bmj.com/


phageal transection to injection sclerotherapy.59 61The first of these trials compared sclerotherapy(immediate and long-term) to stapling transection(with or without splenectomy) in 76 patients; only aminority were actively bleeding at the time oftreatment.59 Rebleeding during the hospital admis-sion occurred significantly less often in the tran-sected patients (3 and 49%, P< 0.01) although themortality for the same period was similar (24 and33% respectively). The second trial divided 70patients with ongoing haemorrhage into good andpoor risk groups on the basis of the severity of liverdisease; in the former, stapling oesophageal tran-section (without devascularization) was comparedto shunt surgery and, in the latter, to injectionsclerotherapy (immediate and long-tbrm).6' Oeso-phageal transection controlled bleeding in 79% ofgood risk and 71% of poor risk patients: this wasless than observed with shunt surgery (100%) andsimilar to that observed with sclerotherapy (83%).Rebleeding occurred in 26 and 43% of the tran-sected patients which was more than observed aftera shunt (13%) but similar to that observed follow-ing sclerotherapy (50%). Severe encephalopathy in15% of the good risk group following shuntsurgery was considered a major limitation. Therewere no differences between the comparativegroups with respect to mortality. The most recentand largest trial compared stapling transection(without devascularization) to injection sclero-therapy (single session) in 101 patients (56% GradeC) the majority with active bleeding at the time oftreatment.6' This was controlled in 88% of thosetransected and 82% of the sclerotherapy treatedpatients. Rebleeding during the period of admis-sion was significantly less frequent following tran-section (7 and 42%, P <0.01). Admission mortal-ity was similar (35 and 44%).

These data support the efficacy of staplingtransection for the management of an episode ofvariceal bleeding. Previous concerns that transec-tion without devascularization might be associatedwith a high rate of rebleeding remain largelyunsubstantiated. However, a beneficial effect uponrebleeding represents the only major advantageover sclerotherapy. The possible adverse effects ofabdominal surgery for subsequent liver transplan-tation have been referred to above.

A strategy for the management of an episode ofvanceal haemorrhage

The diagnostic endoscopy in a patient suspected ofbleeding from varices should be timed to followinitial restoration of the circulating volume(usually within 4 hours of admission). Such earlyendoscopy is likely to improve diagnostic accuracyand also to minimize the delay in starting specific

treatment. The diagnostic endoscopy also providesthe opportunity to carry out immediate sclero-therapy in both those who continue to bleed andthose in whom this has stopped spontaneously. Theefficacy of this approach both for the control ofactive bleeding and for the prevention of earlyrebleeding, favours immediate sclerotherapy as theoptimum initial management.

In the absence of the appropriate endoscopicexpertise the therapeutic options lie between vaso-constrictor therapy and balloon tamponade. It isunlikely that any vasoconstrictor regime will in-fluence severe haemorrhage (sufficient to preventblood volume restitution) and balloon tamponademay be life-saving in this situation. There are fewother indications for this unpleasant and hazar-dous treatment. The choice of a vasoconstrictorregime has been influenced by recent evidencedocumenting the poor results with vasopressin as asingle agent. There is now considerable evidence tosupport the addition of nitroglycerin to vasopres-sin or glypressin therapy. There is no evidence tosuggest that somatostatin has any advantage overvasopressin and nitroglycerin and therefore priceshould caution against its use. Vasoconstrictortherapy and balloon tamponade are temporarymeasures which should be used for short periods(6-12 hours) to obtain initial haemostasis. Failureto control bleeding over a 6 hour period is anindication to change therapy.5 Temporary measuresshould only be discontinued when more definitivetherapy has been instigated.

Injection sclerotherapy is the most widely usedtreatment to prevent early rebleeding althoughthere is no evidence to suggest that this is superiorto stapling oesophageal transection, particularly inGrade A and B patients. The failure to controlbleeding by a single session ofsclerotherapy usuallyreflects incomplete treatment or an incorrect diag-nosis of the bleeding point. In such circumstances afull endoscopic examination should exclude otherbleeding points before repeating sclerotherapy.Failure to control bleeding by two consecutivesessions of sclerotherapy is an indication to changetherapy. Stapling oesophageal transection (with orwithout devascularization) and several types ofportal systemic shunt have been used with successas 'rescue' meaures; the decision between theseoptions is based upon several factors including theavailable expertise, the severity of liver disease andthe possibility of future liver transplantation.The endoscopic diagnosis of bleeding arising

from fundal varices requires an alternative stra-tegy. Attempts at sclerotherapy during activebleeding are seldom successful and the admissionmortality considerably higher than observed inpatients bleeding from oesophageal varices. Sur-gical intervention almost certainly represents theoptimum approach; the options are between devas-

copyright. on M


http://pmj.bm

j.com/

Postgrad M


ebruary 1991. Dow

nloaded from

http://pmj.bmj.com/


cularization or a portal systemic shunt. Splenec-tomy alone is the operation of choice in patientswith splenic vein thrombosis and isolated fundalgastric varices. In patients considered unfit for

surgery measures such as vasoconstrictor therapyor balloon tamponade may provide a bleed-freeinterval during which tissue adhesive can beinjected into the fundal varix.

References

1. Burroughs, A., D'Heygere, F. & McIntyre, N. Pitfalls instudies of prophylactic therapy for variceal bleeding incirrhotics. Hepatology 1986, 6: 1407-1413.

2. Graham, D. & Smith, J. The course of patients after varicealhemorrhage. Gastroenterology 1981, 80: 800-809.

3. Olsson, R. The natural history of esophageal varices. Aretrospective study of 244 cases with liver cirrhosis. Digestion1972, 6: 65-74.

4. Peterson, W., Barnet, C., Smith, H., Allen, M. & Corbett, D.Routine early endoscopy in upper gastro-intestinal tractbleeding. A randomised controlled trial. N Engl J Med 1981,304: 925-929.

5. Mitchell, K., MacDougall, B., Silk, D. & Williams, R. Aprospective reappraisal of emergency endoscopy in patientswith portal hypertension. Scand J Gastroenterol 1982, 17:965-968.

6. Moreau, R., Roulot, D., Braillon, A. et al. Low dose ofnitroglycerin failed to improve splanchnic hemodynamics inpatients with cirrhosis: evidence for an impaired cardiopul-monary baroreflex function. Hepatology 1989, 10: 93-97.

7. Lautt, W. & Greenway, C. Conceptual review of the hepaticvascular bed. Hepatology 1987, 7: 946-951.

8. Greig, P., Langer, B., Blendis, L., Taylor, B. & Glynn, M.Complications after peritoneovenous shunting for ascites.Am J Surg 1980, 139: 125- 13 1.

9. Post, J., Rose, J. & Shore, S. Intravenous use of salt-pooralbumin: effects in 34 patients with decompensated hepaticcirrhosis. Arch Intern Med 1951, 87: 775-778.

10. Kravetz, D., Sikuler, E. & Groszmann, R. Splanchnic andsystemic hemodynamics in portal hypertensive rats duringhemorrhage and blood volume restitution. Gastroenterology1986, 90: 1232-1240.

11. Kravetz, D., Bosch, J., Arderiu, M., Pizcueta, P. & Rodes, J.Hemodynamic effects ofblood volume restitution in rats withportal hypertension due to cirrhosis of the liver: influence ofthe extent of portal systemic shunting. Hepatology 1989, 9:808-814.

12. Panos, M., Moore, K., Vlavianos, P. et al. Sequentialhemodynamic changes during single total paracentesis, andright atrial size in patients with tense ascites. Hepatology1990, 11: 662-667.

13. Merigan, T., Plotkin, G. & Davidson, C. Effect of intra-venously administered posterior pituitary extract on hemorr-hage from esophageal varices. N Engl J Med 1962, 266:565-569.

14. Conn, H., Ramsby, G., Storer, E. et al. Intra-arterialvasopressin in the treatment of upper gastrointestinal bleed-ing: a prospective, controlled clinical trial. Gastroenterology1975, 68: 211-221.

15. Mallory, A., Schaefer, J., Cohen, J., Holt, S. & Norton, L.Selective intra-arterial vasopressin infusion for upper gastro-intestinal tract hemorrhage. Arch Surg 1980, 115: 30-32.

16. Fogel, M., Knauer, M., Kemeny, J. et al. Continuousvasopressin in active upper gastro-intestinal bleeding. AnnIntern Med 1982, 96: 565- 569.

17. Freeman, J., Lishman, A., Cobden, I. & Record, C. Control-led trial of terlipressin (glypressin) versus vasopressin in theearly treatment of oesophageal varices. Lancet, 1982, ii:66-68.

18. Kravetz, D., Bosch, J., Teres, J., Bruix, J., Rimola, A. &Rodes, J. Comparison of intravenous somatostatin andvasopressin infusions in treatment of acute variceal haemorr-hage. Hepatology, 1984, 4: 442-446.

19. Freeman, J., Cobden, I. & Record, C. Placebo controlled trialof glypressin in the management of acute variceal bleeding. JClin Gastroenterol 1989, 11: 58-60.

20. Conn, H. A plethora of therapies. In: Westaby, D., Mac-Dougall, B. & Williams, R. (eds) Variceal Bleeding. Pitman,London, 1982, pp. 221-252.

21. Groszmann, R., Kravetz, D., Bosch, J. et al. Nitroglycerinimproves hemodynamic response to vasopressin in portalhypertension. Hepatology 1982, 2: 757-762.

22. Mols, P., Hallemens, R., Van Kuyk, M. et al. Hemodynamiceffects of vasopressin, alone and in combination with nitro-prusside in patients with liver cirrhosis and portal hyperten-sion. Ann Surg 1984, 199: 176-181.

23. Westaby, D., Gimson, A., Hayes, P. & Williams, R. Haemo-dynamic response to intravenous vasopressin and nitro-glycerin in portal hypertension. Gut 1988, 29: 372-377.

24. Gimson, A., Westaby, D., Hegarty, J., Watson, A. &Williams, R. A randomized trial ofvasopressin and vasopres-sin plus nitroglycerin in the control of acute variceal hemorr-hage. Hepatology 1986, 6: 410-413.

25. Tsai, Y., Lay, K., Ng, W. et al. Controlled trial ofvasopressinplus nitroglycerin vs vasopressin alone in the treatment ofbleeding esophageal varices. Hepatology 1986, 6: 406-409.

26. Bosch, J., Groszmann, R., Garcia-Pagan, J. et al. Associationof transdermal nitroglycerin to vasopressin infusion in thetreatment of variceal hemorrhage: a placebo-controlledclinical trial. Hepatology 1989, 10: 962-968.

27. Cort, J., Albrect, I., Novakova, J., Muller, J. & Jost, K.Regional and systemic haemodynamic effects of some vaso-pressins. Structural features of the hormones which prolongactivity. Eur J Clin Invest 1975, 5: 165-175.

28. Vosmik, J., Jedlicka, K., Mulder, J. & Cort, J. Action of thetriglycyl hormonogen of vasopressin (glypressin) in patientswith liver cirrhosis and bleeding esophageal varices. Gastro-enterology 1977, 72: 605-609.

29. Walker, S., Stiehl, A., Raedsch, R. & Kommerell, B. Terli-pressin in bleeding esophageal varices, a placebo-controlled,double-blind study. Hepatology 1986, 6: 112-115.

30. Lin, H., Tsai, Y., Lee, Y. et al. Systemic and portalhaemodynamic changes following triglycyllysine vasopressinplus nitroglycerin administration in patients with hepatitisB-related cirrhosis. J Hepatol 1990, 10: 370-374.

31. Fort, E., Sautereau, D., Silvain, C., Ingrand, P., Pillegrand, B.& Beauchant, M. A randomised trial of terlipressin plusnitroglycerin vs Balloon tamponade in the control of acutevariceal hemorrhage. Hepatology 1990, 11: 678-681.

32. Bosch, J., Kravetz, D. & Rodes, J. Effects of somatostatin onhepatic and systemic hemodynamics in patients with cirrhosisof the liver, comparison with vasopressin. Gastroenterology1981, 80: 518-525.

33. Jenkins, S., Baxter, J., Corbett, W., Dewitt, P., Ware, J. &Shields, R. A prospective randomised controlled clinical trialcomparing somatostatin and vasopressin in controlling acutevariceal haemorrhage. Br Med J 1985, 290: 275-278.

copyright. on M


http://pmj.bm

j.com/

Postgrad M


ebruary 1991. Dow

nloaded from

http://pmj.bmj.com/


34. Burroughs, A., Sprengers, D., Hughs, M., D'Heygere, F.,McCormick, A. & McIntyre, N. Randomised, double-blindplacebo controlled trial of somatostatin for control ofvariceal bleeding. J Hepatol 1988, 7 (Suppl 1): S14 (abstract).

35. Valenzuela, J., Schubert, T., Fogel, R. et al. and a multicenterstudy group. A multicenter randomized, double-blind trial ofsomatostatin in the management of acute hemorrhage fromesophageal varices. Hepatology 1989, 10: 958-961.

36. Mastai, R., Grande, L., Bosch, J. et al. Effects of metoclo-pramide and domperidone on azygos venous blood flow inpatients with cirrhosis and portal hypertension. Hepatology1986, 6: 1244-1247.

37. Hosking, S., Doss, W., El-Zeiny, H., Robinson, P., Barsoum,M. & Johnson, A. Pharmacological constriction of the loweroesophageal sphincter, a simple method of arresting varicealhaemorrhage. Gut 1988, 28: 1098-1102.

38. Rowntree, L., Zimmerman, E., Todd, M. & Ajac, J. Intra-esophageal venous tamponage. Its use in a case of varicealhemorrhage from the esophagus. JAMA 1947, 135: 630-63 1.

39. Vlavianos, P., Gimson, A., Westaby, D. & Williams, R.Balloon tamponade in variceal bleeding: use and misuse. BrMed J 1989, 298: 1158.

40. Mitchell, K., Silk, D. & Williams, R. Prospective comparisonof two Sengstaken tubes in the management of patients withvariceal haemorrhage. Gut 1980, 21: 570-573.

41. Teres, J., Cecilia, A., Panes, J., Rimola, A., Bru, C. & Rodes,J. Esophageal tamponade for bleeding varices. Controlledtrial between Sengstaken-Blakemore Tube and the Linton-Nachlas Tube. Gastroenterology 1978, 75: 566- 569.

42. Sarin, S. & Nandy, S. Balloon tamponade in the managementof bleeding oesophageal varices. Ann R Coll Surg Engl 1984,66: 30-32.

43. Johnston, G. & Rodgers, H. A review of 15 years experiencein the use of sclerotherapy in the control of acute haemorr-hage from oesophageal varices. Br J Surg 1973, 60: 797- 800.

44. Terblanche, J., Yakoob, H., Bornman, P. et al. Acutebleeding varices. A 5 year prospective evaluation of tampon-ade and sclerotherapy. Ann Surg 1981, 194: 521-530.

45. Paquet, K.-J. & Feussner, H. Endoscopic sclerosis andoesophageal tamponade in acute hemorrhage from eso-phago-gastric varices: a prospective controlled randomisedtrial. Hepatology 1985, 5: 580-583.

46. Yassin, Y. & Sherif, S. Randomised controlled trial ofinjection sclerotherapy for bleeding oesophageal varices - aninterim report. Br J Surg 1983, 70: 20-22.

47. Larson, A., Cohen, H., Zweiban, B. et al. Acute esophagealsclerotherapy. JAMA 1986, 255: 497- 500.

48. Soderland, C. & Ihre, T. Endoscopic sclerotherapy versusconservative management of bleeding oesophageal varices.Acta Chir Scand 1985, 151: 449-456.

49. Moreto, M., Zaballa, M., Bernal, A., Ibanez, S., Ojembar-rena, E. & Rodriguez, A. A randomised trial oftamponade orsclerotherapy as immediate treatment for bleeding eso-phageal varices. Surg Gynecol Obstet 1988, 167: 331-334.

50. Westaby, D., Hayes, P., Gimson, A., Polson, R. & Williams,R. Controlled trial of injection sclerotherapy for activevariceal bleeding. Hepatology 1989, 9: 274-277.

51. Hosking, S. & Johnson, A. Gastric varices, a proposedclassification leading to management. Br J Surg 1988, 75:747-750.

52. Gimson, A., Westaby, D. & Williams, R. Endoscopic therapyin the management of gastric variceal haemorrhage. JHepatol 1990 (in press).

53. Soehendra, N., Grimm, H., Nam, V. & Berger, B. N-Butyl-2-cyanoacrylate, a supplement to endoscopic sclerotherapy.Endoscopy 1987, 19: 221-224.

54. Van Stiegman, G. & Goff, J. Endoscopic esophageal varixligation, preliminary clinical experience. Gastrointest Endosc1988, 34: 113-117.

55. Orloff, M., Bell, R., Hyde, P. & Skivolocki, W. Long-termresults of emergency portacaval shunt for bleeding esopha-geal varices in unselected patients with alcoholic cirrhosis.Ann Surg 1980, 192: 325-340.

56. Cello, J., Grendall, J., Crass, R., Weber, T. & Trunkey, D.Endoscopic sclerotherapy versus portacaval shunt in patientswith severe cirrhosis and acute variceal hemorrhage, long-term follow up. N EnglJ Med 1987, 316: 11-15.

57. Warren, D., Henderson, M., Millikan, W. et al. Distalsplenorenal shunt versus endoscopic sclerotherapy for thelong-term management of variceal bleeding. Ann Surg 1986,203: 454-461.

58. Shaw, B., Wood, P., Gordon, R., Iwatzuki, S., Gillquist, W.& Starzl, T. Influence of selected patient variables andoperative blood loss on 6 month survival following livertransplantation. Semin Liver Dis 1985, 5: 385-393.

59. Huizinga, W., Angorn, 1. & Baker, L. Esophageal transectionversus injection sclerotherapy in the management of bleedingesophageal varices in patients at high risk. Surg GynecolObstet 1985, 160: 539-546.

60. Teres, J., Baroni, R., Bordas, J., Visa, J., Pera, C. & Rodes, J.Randomised trial of portacaval shunt, stapling transectionand endoscopic sclerotherapy in uncontrolled variceal bleed-ing. J Hepatol 1987, 4: 159-167.

61. Burroughs, A., Hamilton, G., Phillips, A., Mezzanotte, G.,McIntyre, N. & Hobbs, K. A comparison of sclerotherapywith staple transection of the esophagus for the emergencycontrol of bleeding from esophageal varices. N Engl J Med1989, 321: 857-862.

copyright. on M


http://pmj.bm

j.com/

Postgrad M


ebruary 1991. Dow

nloaded from

http://pmj.bmj.com/

article the - postgraduate medical journal142 a.e. gimson&d. westaby rate in controlling...

Documents