Therapeutic options for patients bleeding with peptic ulcers

ABR THOMSON. Therapeutic options for patients bleeding with peptic ulcers. Can J Gastroenterol 1994;8(4):269-274. lt is likely that the besc outcome for the patient with an acute upper gastrointestinal bleed (GIB) includes early d iagnosis: for a bleeding lesion with a high risk of rebleeding, in an older patient with systo lic h.ypotension or in a person with multiple medical problems. Early therapeutic endoscopy with meticulous control of inrragastric pH will Likely achieve the best outcome. The ideal pH criterion to stop bleedmg or to prevent recurrence is unknown. A n algorithm is presented to guide the clinical management of patiencs with GIB, and to focus on important questi.ons for future therapeutic studies.

U PPER GASTROINTESTINAL TRACT bleed (G IB) is a common med ical emergency which constitutes approxi-maLe ly 15% of the workload of major Canadian teaching universities (1 ).About 30% of these patients rebleed, 20% require emergency surgery and 5% die (2,3 ).Surgery is usually indicated for severe, persistent or recurrent bleeding.Any treatment chat might Jclay surgery to optimize the oppo rtu ni ty to stabilize patients, or prevent its need in the first place, would be useful.An y med ical t herapy cha t could reduce the risk of rebleeding o r death would be of even greater value.
The therapeut ic options for patients with bleedi ng peptic ulcers include surgery, therapeutic endoscopy (Table 1) and drugs.After appropriate resuscitation, current medical therapy incl udes mechanical compression of the bleeding vessel using the various fo rms of therapeutic endoscopy.The choice of drugs that might improve the clinical outcome of patients with GIB includes antifibrinolytic therapy, reduction in splanchnic pressure with somatostati n (or a long acting analogue such as octreotide) or acid-lowering measures using H z receptor antagonists or proton pump inhibirors.

THERAPEUTIC ENDOSCOPY
Diagnostic enJoscopy has assllmed an 1mptirtanc role 111 rhe management of patients with GIR hecallse of its accllraq 111 esrnhlishing a diagnosb, in de-term111111g which patient 1s likely 10 concmue bleeding and in identifymg which patient may henefit from ,1 therapeutic mtervention (4).From the reported resu lts of controlled clinical trials ( 5,6), participants of the 1989 United States Nauonal Institute ~if l lealth Comensus Conference (7) recommended th,H therapeutic endoscopy he the hemostat ic procedure of fin,t Lhoice for pauents \\ ith ulcers at n,k of rebleedmg.Numemus forms of endoscopy therapy are avai lable (Table I).Sacks and culleagues (6) performed a meta-analysis that demonst ri'ltc<l that therapeutic en<losrnp) reduces the need for emergency surgery hy nenrly two-thirds, and reduces the mortality rate of patients with t,IB by nearly nne-thir<l.Injection therapy is equal to that of thermal methods such as YAG laser (8) or mult ipolar cleccrocoagulation (9), while equal or inferior results are achieved wnh heat probes ( I 0, 11 ).
Let lls consider an important study that stnmgly emphas1:es the importance nf ther,1peut1c endoscopy in GIB.In a series of 1880 patients a<lmmed consecutively to a single hospital in Barcelona wnh a Heeding peptic ub:r, H l had a h igh risk nf furt he r hemorrhage as assessed h) the presence of active arteria l hlecding or a nonhle1;:ding \'tsible vessel ( l 2 ).These were older patients with ,1 mean age of 64.9±15 years.The location of the ulcers was: 58% duodenal, 32.3 1 \i gastric, 5 1 \i Momal and 4.7% pyloric.A nonblced-111g v1s1hle vessel was seen 111 65.1 °11, oozing 111 3 1.4% and spurting in 3.5%.A<lrenaline w.,s injected ( I per 10,000, aliquots of 1.0 to 2.0 mL first around ranitidine 20
Unfonunmcly, the use of t ranex• anm: aud 1s cnmplic,11ed hy an 111• cre,hed 111c1dentT of rhromhnphleh1t1s  ( 19).OespiLe Lhis, iL remains unclear why this form of therapy has not gaineJ wider acceptance.The potential mechanism of benefiL from this therapeutic agem is also uncertain.

SOMATOSTATIN AND OCTREOTIDE
Agents reducing splanchnic blooJflow have been stuJieJ in the treatment of patients presenting with hematemesis anJ melena.The hormone somacostatin reduces gastric aciJ secretion (20), pepsin (21) and gastrin secretion (22), and in noncirrhotics iL reJuces splanchnic blood/low (23 ).A l.1rge J ouhle-blind comrol led single cencre stuc.ly was performed in 630 unselected patients admined to the Notti ngham hospitals wiLh hematemesis and melena (24).Placebo or somato-~tatin 250 µ g was given in an initial intravenous bolus followed by a 72 h intravenous saline infus ion providing 250 µ g of somarosrntin or placebo hourly.No difference was noted between somaLOstatin versus placebo in terms of rebleecling -70 of 3 l5 (22%) with somatosrntin versus 89 of 315 (28%) with placebo -or operation rates -35 of 3 l5 ( l I%) with somatostatin versus 25 of 315 (8%) with placebo.None o( these differences was statisticall y sign ificant.H owever, it should be noted that 'all comers' were accepted and it could be argued that the inclusio n c riteria should be limited to those groups of patients in whom a low placebo-response rare can be expected to a llow the trial drug a chance to show its poLential efficacy. In the study of somatostatin performed by Langman and colleagues (25), a generous dose of somatostatin was used, yet the study was negative.T he somacosratin was given by infusion after a loading Jose, so chm t he short hal f-l ife of aboUL 3 mins is not a consideration.The somatostatin ana logue ocrreoride (SMS 20 1-995, Sandoz, Basel, Switzerland) has a half-life of about 100 mins.Ln a second study, patients were included with GIB from bleeding gastric or Juodenal ulcers requiring CAN J GASTROENTEROL VOL 8 NO 4 jULY/AUGU:, 'T l 994 blood transfusion or plasma expanders.C hristiansen et a l (26) conducted a controlled 23 multicencre study in Denmark and Germany, with 123 patients on placebo and 115 o n octreotide.There was no difference in the rates of stopping bleeding and preventing rebleeding between placebo (70.6%) a nd ocrreo tide (69.6%), nor any difference between the two groups in surgery rates, blood transfusion requirements a nd time required before bleeding stopped.
Vasopressin has been used to treat patients with nonvariceal upper GIB, but there may be serious complications in those with cardiovascu lar or renal disease, a nd efficacy is marginal (27-29).

HISTAMINE H2 ANTAGONISTS
U ntil Septembe r L 985 there were 2  This open stud y success has been accompanied by a numher of brief reports of the beneficial use of o mepra-:olc ro control h e mo rrhage from acute h e mo rrhagic gasrritis (33 ), in five c rit ically ill patients with life-threatening n on vari cca l uppe r GIB who h ad fa iled to respond ro conventio nal therapy (34) a n<l in 11 o lder pat ients not responding to raniridinc or somarostatin, with persistence of clinica l or endoscopic s igns nf bleeding due to e rosive gastritis or acute peptic ulcer (the hemoglobin con centration was nor raised despite four units of blood) (35).
The positive results of the study of Bruner a nd C hang (32) may h ave been due co the inclusion of patients with bleedi ng (36).The presence of active arterial bleeding at e me rge ncy e ndoscopy is associated wi th furthe r h e morrhage in 95% of cases ( 4), whi le a nonhlecding visibl e vessel on the ulcer Ooor resu lts in a rc bleeding rare of33 to 55%. S tigmata of bleedi ng also pmvides a ta rget ro whic h e ndoscopic therapy must be addressed to obliterate the unde rl ying a rte ry ( 14,37).What is the ra tio n ale for using acid lowe ri ng therapy to tre::it patients with GIB? Humeral-and thrombocyte-i nduced hcmosrnsis o nl y occurs at pH values above 6.0 (38).Thromhocyte aggregation and clot forma ti on a rc inhi bited in th e presence of even sm all a mounts of acid, and n ewly formed c lots a re subject to rapid digestion by gastric juice at pH va lues below 5.0 (39).Pepsin fur th er e nhances platelet disaggregation.This argues for meticulous imragastric pl-I control in ratients with upper gastrointestina l bleeding.
The vari ous therapeutic regimens used in these studi es include l-12 receptor antagonists g ive n hy bolus, infusion o r bolus plus in fusio n , a nd omeprazolc given by bolus.The exten t of acid cont rol (pH greate r th a n 5.0) is better with c imc tidine by infusio n versus hy bo lus (40) a nd a loading dose of famotidine achieves supe ri o r ac id control when given before infusion, compared with infusion alone (unpublish ed data ).A load ing dose o f rani ti<l ine docs not add co th e acid inhibiting effect of a n in fi.1sion o r ranitidine (unpublish ed data ).The infusion of ra nitidine results in considerable d iurna l rhythm of intragastric va lues of pH, with great varia tion between subjects (Figure 2).Bleeding peptic ulcers 42) (Figure 3).It now becomes appropriate to test this ne w a nd apparently idea l regimen in a clinical setting.
T hree trials repo rted in abstract fo rm h ave suggested that there may be a lower rate of reblceding in patients with peptic lesions no t due co nonsteroidal a n ti-infl ammatory drugs, in whom anti-hclicobactcr therapy was initiat ed at the t ime of diagnosis of a b leed ing peptic ulce r.Inc reasing evidence is accu mulating to buttress the appropria ten ess of eradicctting /-lelicobacier pylori LO reduce rhe frequency of ulcer relapse, a nd this erad ication therapy appears co be appropriate in patients presenting with a bleeding peptic lesion.
In summary, rh e patient w ho presen ts with acute uppe r GIB needs prompt cli nical assessment a nd resusc itati on (Figu re 4).Early e ndoscopy is required for diagn osis therapy.O nl y patients with act ive bleed ing or high risk o f rebl eecling, or medically e ndangered pa tie nts require admission to h ospital.The role of routine use of systemic infusions of H z receptor antagonists is not prove n, a nd the prom ising ro le for meticulo us contro l of inrragastric pl-I using proton pump inhibitors n eeds to be establi shed.
nificant prognostic factors, and the presence of a gastric or duode na l ulcer w as a significanL foc tor for rehleeding and operatio n but nm fo r death.No strat ification was d o ne fo r s ire or initial activity of hl eeding lesio ns.In cont rast to th is negati ve placehnconrrolled study with 'all comers', Brunne r a nd C hang(32) performed a sma ller open rando mi:ed cont rolled trial (Figure l ) comparing intravennu, therapy with ran itidinc or rnnepra:olc (o meprazole by bolus, 80 mg initially fo llo wed hy 40 mg every 12 h for five days ver:;us rn111tidme SO mg fo llowed hy cont inuo us infus ion of 400 mg rn111t idinc pe r day ).Pat ien ts we re c riLicall y ill with act ivel y hlced ing pept ic ulceration.S uccessful treatment was proven hy follo w-up endoscopy o n d ay 6, unle~, there w,is earlier treatme nt fa ilure.Of 20 patie nti> in t he ranitidinc group, 17 were treatment fa ilures with more th an 2.5 L of hl ood necessary tn mai ntain ,1 he moglobin level ahm•c 10 g/L (Figure I ).Bleeding could he controlled suhsequently in l3 of 17 pm ie nts after c ha nging to o meprazole (76% ), a nd 16 nf 19 pat ients started on n meprazolc were successfull y trenrcd (84% ).