Review of endoscopic thermal treattnent of peptic ulcer hemorrhage

Peptic ulcer hemorrhage is still an important cause of emergency surgery and death. The overall mortality is around 10% from gastrointestinal bleeding, and most of the preventable deaths occur in elderly patients with continued or recurrent bleeding from peptic ulcers. An effective nonsurgical method of hemostasis has long been recognized to be desirable. However it was only when the 'visible vessel' was recognized as the important risk factor for further bleeding that studies capable of testing new modalities adequately could be performed. Careful wa:;hing of the ulcer crater is essential for identification of ,hese visible vessels. An effective endoscopic method was first demonstrated in 1981 in patients with visible vessels treated with argon laser. Many groups have now shown excellent efficacy of neodymium:yttrium-aluminum-garnet (NdY AG) laser in preventing further hemorrhage from ulcers with bleeding and nonbleeding visible vessels. Two controlled prospective studies have demonstrated efficacy of the heater probe, but one well designed study did not. Similar studies with both bipolar and monopolar electrocoagulation have shown significant reductions in rebleeding in patients with visible vessels treated using the chosen modality. More recent studies have achieved excellent results by pre-injection with adrenaline and one repeat endoscopic treatment for rebleeds. A few groups have now reported equally good results with injection alone. Long term follow-up of patients with peptic ulcer hemorrhage has confirmed prolonged hemostasis in groups treated with two thermal modalities and in controls. Can J Gastroenterol 1990;4(9):653-662

remains an important cause of hospital admission, emergency su rgery and death.In the United Kingdom, upper gasrrointestmal tract bleeding 1s responsible for 50 to 100 acute admissions per 100,000 population per year (1,2).Peptic ulcers account for 50 to 60% of t hese cases.The mortality from upper gastrointestinal trace hemorrhage has remained largely unchanged over the past 30 years at around 10% (1,3).This b despite improvements in resuscitation, intensive care, new surgical techniques and the introduction of fibreoptic endoscopy, probably due to the ever increa.~ingproportion of elderly patients (3,4 ).
In a comprehensive analysis of 484 patients with gastrointestinal bleeding (3),itwasfoundthatof55(1 l.4%)who died, 22 (40%} were potent ially avoidable deaths (either postoperative complications or persistent bleeding).The majority of patients who d ied had bleeding peptic ulcers, and all but one were 60 years of age or older.The study did not comment on the relative risk of rebleeding, but it has previously been documented that further peptic ulcer hemorrhage during hospital admission is associated with a l 0-co l 2-fold increase in mortality (5).
improve overall mortality rate.The first, which had been previously advocated (I), was to perform early surgery so that patients would be in better condition to withstand the operation.The second was to adopt a more conservative surgical approach in the elderly in the hope that some patents would stop bleeding spontaneously.It was also suggested th t the most promising means of reducing mortality would be a nonsurgical method of cont roll mg bleeding so that emergency surgery could be avoided (3 ).
Groups advocating early surgery have reported excellent results (6,7) with overall mortality less than 4%, bur to achieve this, operation rates of 30 to 60% were required.The only controlled comparison of early versus late surgery was performed in 1984 (8).In the elderly group, early surgical intervention significantly reduced mortality, but only in patients with bleeding gastric ulcers.The price for t his improvement in mortality was a 62% operation rate in the early surgery group compared with 27% in the delayed surgery group.S ince these trials, the prognostic accuracy of diagnostic endoscopy has improved.
It is now known that the detection of a 'visible vessel' at endoscopy constitutes a high risk of fu rther bleeding.It is likely therefore that a policy of early surgery in combination with careful diagnosuc endoscopy would result m fewer unnecessary operations.Such an approach in this mcreasmgly e lderly group is still likely to result in significant morbidity and mortality.Fortunately, endoscopic techniques have progressed and can now provide the nonsurgical method advocated in 1979 (3).The majority of patients are now candidates for some kind of endoscopic therapy capable of terminating bleeding.Before considering trials of endoscopic hemostasis for peptic ulcer hemorrhage, it is important to consider identification of ulcers at high risk of rebleed ing as it is only by including these in studies that hemostat1c modalities can be adequately assessed.

ENDOSCOPIC PREDICTION OF FURTHER BLEEDING
C linical indicators provide a guide to the likely outcome of a bleed in an individual.It has long been known that important negative prognostic factors on admission are: low blood pressure, gross anemia, old age and further bleeding (I).Endoscopic indicators have, however, proven more useful.Stigmata of recent hemorrhage in peptic ulce rs that had bled were first identified and found to be associated with increased risk of further bleeding in 1978 (9).Three stigmata were described: fresh hleeJing from the lesion; fresh or a ltered blood c lot or black slough adherent to the lesion; and a vessel protruding from the base or margin of the ulcer.
In this earl y study 2 5 of 60 ulcers (42%) with stigmata of recent hemor, rhage reblcd compared with one of 29 (3%) without stigmata.In a pro~pecnl'e tnal in 1981 (1 O) it was reported that when the ulcer crater was washed at endoscopy to clear overlying blood and secretions, visible vessels were founJ m 56 of 11 7 patients ( 48%) where full examination of the ulcer crater was pos, siblc.This 1s a higher propornon d patients than previously reported.Jn th e absence of a spurting vessel or pulsating pseudoaneurysm, the visible 1e;sel was defined more fully as "a rcJ IJ! blue spot, resistant to gcnrk washing, often associated with red clot and ~,.most a lways unique in the ulcer crater.
The authors reported that the presence of a visible vessel implied a far Most se ries have confirmed the predictive value of the visible 1•t -.el (13-15).There are, however, two recent prospective stud 1es in which thr visible vessel did not emerge as a su~rior predictor of further hemorrha~ie (16,17).In the fi rst, ulcers were n,x washed, and therefore many v1s1ble ves, sels must have been missed.In the second, although ulcers were washc<l,a proportion of ulcers remained covereJ with adhe rent clot and were therefore inadequately characterized .In two studies ( 16,18) the ass0<.iation of ~hod on admission with a visible vessel or an adherent clot at endoscopy was founJ to be a bener p re<l1crnr of rebleeding than en<loscopic signs alone.

NATURE OF T HE VISIBLE VESSEL
In 1979 it was suggesteJ that the visible vessel was an acutely erodeJ artery standmg up m the floor of the ulcer ( 19).However a pathologic study of gastric ulcers with visible vessels which were subsequently resected (20) demonstrated that the artery only protru<leJ above the ulcer crater m one in five cases.The v1s1ble vessel in the majority was in fact a thrombus extenJmg from a rent in the s1Je of the artery.The alternative terms 'sentinel clot' or 'pseudoaneurysm' are therefore more appropriate.

RES UL TS OF T RIALS
There have been many anec<lmal reports of successful laser anJ other thermal treatments of upper gastrointestinal tract hemorrhage; however, as the vast majority of these blee<ls cease spontaneously, true assessment of efficacy can only come from looking at patients known to be at h igh risk of rLbleeding in controlled and comparative stu<lies.Stu<lies of thermal methods to control peptic ulcer hemorrhage will now be considered.In all reports the inc idence of rebleeding in ulcers without v1s1ble vessels has been low.Characterization of the ulcer crater fo llowing adequate washing is a common difference between studies which demonstrate benefit from the various techniques and similar studies which do not.Argon laser: Two early studies ( 14,21) showed no significant difference in rebleed mg rate, emergency surgery or mortality between ulcers treated with argon laser and control ulcers.In the first study there was a trend towards reduction in rebleeds in patients with spurting vessels, wh ich d id not reach significance.The authors of this study commenced that their resu lts required ampltficatton m larger trials.In a twocentre trial ( 11 ), further hemorrhage occurred in significantly fewer patients with v isible vessels treated with laser compared co the control group.This trial demonstrated for the first time a Endosc opic thermal treatment NdY AG laser: This laser has greater power output and tissue penetration and supenor hemostanc efficiency in animal studies than the argon laser (22).Laser output 1s focused onto the end of a 400 co 600 µm quartz fihre.This fibre is contained in a Teflon catheter 2.6 mm m diameter, which gives It mechanical strength and thus a llows it to he passed down the working channel tif an endoscope.It also a llows a coaxrnl stream of gas ro he passed co keep the metal tip of the fibre cool and cleanalso to clear the target of blood and debm.This technique is noncontacc.
The tip of the fihre is positioned 5 to I 0 mm from the target.Pulses arc usually applied circumferentially around the visible vessel in 0.5 to 1 s shots using a power setting of 70 to 80 W.
Three early tna ls (2 3-2 5) showed no benefit of Y AG laser.None of these studies, however, defined the preuse nature of the bleeding point, anJ numbers were small.In aJdinon, one study ( 23) was performed using a power of 50 W, which 1s suboptimal.
A later study (26) demonstrated initial hemnstasis in patients with active norn,purting hemorrhage from created peptic ulcers co be significantly better rhan controls.(Bleeding stopped 111 all    2. Similar results were reported in another study (27) on patients with spurting and nonbleeding visible vessels.There was a trend towards reduction of rebleed ing and the need for emergency surgery which was not statistically significant, bur there was a significant reduction in mortality.
A derai led study from this unit produced convincing results (12).One hundred and thirty-eight patients with peptic ulcer hemorrhage who were found to have stigmata at endoscopy were randomized to receive YAG laser photocoagu la tion or conservative management.Their results are shown in Table 3. Patients with spurting hemorrhage or nonbleeding visible vessels treated with laser had significantly less chance of further hemorrhage than controls.Stmistical significance was shown whether these groups were considered individually, together or as part of the group showing stigmata of recent hemorrhage (Table 3).The incidence of rebleeding in ulcers demonstrating minor stigmata was very low.Ulcers with overlying clot demonstrated intermediate risk of rebleeding, presumably because this group included some patients with hidden visible vessels and some with minor stigmata.The need for emergency surgery and mortality were also significantly reduced in the laser group.The data on relative risk of rebleeding of the various stigmata previously discussed were derived from the combined results of this study and the argon laser study from the same group.
The protocol of chis study was 1dent ica l to rhe argon laser study.The hemostatic efficacy of the NdY AG laser can thus be compared d irectly with that of the argon laser.Such a comparison shows the NdY AG laser to be significantly better in the groups with all stigmata and with visible vessels alone.
A second study from the Belgian group (28) looked at the effect of preinjection with l: 10,000 adrenaline prior to Y AG laser treatment in patients with visible vessels (both spurting and nonbleeding) (Table 4 ).In 87% of the 54 patients studied, complete hemostasis was achieved.The results for bleeding visible vessels were particularly impressive: final hemostasis was achieved in 85% resulting in only L5% (four of 26) requiring emergency surgery.There was no control group in this study, but the figures compare very favorably with ocher published series.One might expect a rebleeding rate of around 50 to 60% in such cases overall -far higher than the l3% observed.The authors ascribed the apparent benefit of pre-injection with adrenaline to temporary hemosrnsis, which al-lowed better visual1zarion of the target It is likely, however, that some hencnt is also derived from vasoconstriction in the underlying artery and a consequent reduction in removal of heat from tht area (the 'heat sink effect').
A further study from the Belgian group (29) and one from the present authors' unit (30), both discusscJ in more detail in the next section, havt contributed co the weight of ev1Jcnce confirming the efficacy ofYAG laser in peptic ulcer hemorrhage.
One recent randomized contwlled study (31) reported no benefit from Y AG laser in bleeding peptic ulcers.
Washing of clots at endoscopy was not performed, and thus the hleeJing pomt was not adequately characteri~eJ m most p,niencs.Only 29 visible vc~se~ were detected in 174 patients inclu<led in the study ( 17%).Th is is very low in comparison with most series.The protocol excluded patients "too un• stable to be moveJ to the laser facility.'During the course of che stu<ly thlS amounted to 221 patients -more than the rnral number included.These patients were those with the most hemodynamically significant hlecds, and as such, they were likely to have visible vessels (perhaps explaining the low pick-up rate) and thus most likely to benefit from laser therapy.The very low mortality (I%) seen in this study surely reflects the exclusion of these high risk patients.Heater probe: Experimental studies (32,33) comparing methods of electrocoagulation suggested the hearer probe to be most promising in terms of efficacy and safety.The first of these studies found the heater probe and the h1polar electrode 10 cause full thickness Jamil&" significantly less frequently than the liquid or dry monopolar electnxlcs (20 versus 60%); thus, the risk of perfora, tion was thought to he lower with these electrodes.The bipolar electrode, however, performed less well than the other three in hemostasis.The seconJ study confirmed the heater probe anJ the bipolar electrode to be safest and found the two modalities co be equally efficacious.
The heater prohe was developed in 1978 (34 ).IL was designed to apply pres-sure and heat simultaneously to a bleeding vessel.The probe comprises a hollow aluminum cylinder with an inner hearer coil and an outer coat mg of nonstick Teflon.It also contains a separate thermocouple element in the tip to measure temperature.It CcUl l,e heated to a maximum of 250°C, a temperature which is maintaineJ until a preset amount of energy is delivered.l11ere is also a proximal irrigation port to allow washing of the target even when the probe is applied forcibly to tissue.The technique employed is to apply the inactivated probe around the bleeding point with moderate pressure in order to tarnponade bloodflow 111 the underlying vessel.The probe is then activated and a preset quantity of energy delivered ( usually 30 J ).This procedure 1s then repeated at multiple sites circumferentially around the l,leedmg point.This technique of vessel tarnponade and thennal coagulation has been termed 'coaptive coagu lation' and applies equally to all thermal merhoJs apart from laser, for which the fibre tip is positioned S to l O mm from the target.
Encouraging results were reponed in a pilot study in 198 3 (35) with an early heater probe.Patients al ready anesthetized for emergency operation were treated.Final hemostasis was achieved in 12 of 15 patients with gastric ulcers; only one of 15 required surgery.Experience with duodenal ulcer~ was not as impressive -only two of IO avoided surgery.There were no deaths in the patients with gastric ulcers, but three died in the duodenal ulcer group.The exact nature of bleeding point was not discussed.A similar pilot study (36) also produced encouraging results-12 of 16 patients proved treatable.(The others had large overlying c lots or active hemorrhage.)Bleeding was arrested in all L 2 patients treated, al though lme required a second application.
A retrospective study ( 3 7) of Y AG laser and heater probe in patients with spurting and nonbleeding visible vessels showed a significant difference in final hemostasis in favour of the heater probe group: 19 of 20 (95%) of ulcers treated with heater probe and 24 of 35 (69%) of ulcers treated with Y AG laser stopped bleed ing ( P<0.05 ).Some ranents who rehleJ had second treatments, a protocol which has been used increasingly in subsequent studies.It is interesting to note the very high rate of induced bleeding (29%) in lasertreated patients.This did not occur in the heater probe group and may account for the apparent difference in fina l hemostasis.The diffe rence in surgery and mortality in the two groups did not reach statistical significance.The authors concluded that heater probe was more effective than laser as well as being quicker and safer; however, the efficacy of heater probe has not been shown to be better than laser in any of the subsequent prospective studies.
The first pro~pective randomized comparison of the two modalities was performed in 1987 (30).PHrients with peptic ulcers showing stigmata of recent hemorrhage ( total 14 3) were randomized to receive Y AG laser, heater probe or no endoscopic treatment.The rebleeding rate was significantly lower in the laser group (20%) than controls (42%).Rebleeding in the heater probe group (28%}, however, was not significantly different from controls (although there was a strong trend towards reduced rebleeding}.There was no srntistical difference in deaths in the three groups.Combined analysis with results of a previous study from the same group ( 12) and with an identical protocol but no heater probe arm showed significantly fewer rehleeds and deaths in patients treated with laser rather than heater prol,e.Ulcers with both major and minor stigmata of recent hemorrhage were included in this study.A further analysis of subgroups may provide more information, although the authors did note that randomization was hiased cowards heater prol,e therapy.
Excellent resu lts have subsequently been reported in an open prospective study (38).Fifty patients with visible vessels (28 with spurting hemorrhage) were treated with heater probe.The results arc shown 111 Table 5.In all, 49 of 50 visible vessels (98%) were controlled with treatment, and only seven ( 14%) re bled within one week.Hemostasis was ach ieved in four of these  A recent prospective randomized controlled trial from Scotland (39) has demonstrated efficacy of heater probe against controls.Patients with active blcedmg were randomized to treatment or no treatment.No rcblecding occurred in 20 ratients treated with heater probe, but five of 23 controls rehled (P=0.05 ).Emergency surgery was performed in three of the control patients.Consider111g ulcers w ith visible vessels, seven of seven in the heater probe group d id not rebleed but Lhree of four controls rebled (P<0.04 ).Bipolar clectrocoagulation: The bipolar probe consists of two or more (commonly three) pairs of bipolar electrodes allowing diathenny with tip angulation.It comes in various sizes, early prohes being 2.3 mm in diameter and later ones 3.2 or 3.4 mm in diameter.There is a central irrigation channel.Early power sources delivered 25 W -the lacer units SOW.The technique of vessel coagulation is identical to that employed with the heater probe.In experimental studies, the 3.2 mm probe has proven more efficient at coagu laLing bleeding mesenteric arteries than the 2.3 mm probe (40).
An earl y multicentre study ( 41) demonstrated clinical efficacy of the bipolar 2.3 mm probe.A total of 44 patients with various actively bleeding  Three controlled trials were conducted with the 2.3 mm probe.In the first ( 42), 46 patients with peptic ulcers which had bled were randomized.The ulcers were not characterized further, anJ nod ifferences were shown between the 2 l treated patients and controls.One treateJ patient suffered massive bleeding during the procedure.The seconJ (43) found no difference between 21 treated and 24 control patients with stigmata of recent hemorrhage.Access with the probe, however, was inadequate in half of the patients treated.The third trial with this small probe did show a statistical reduction in rebleeding in patients treated with the 2.3 mm bipolar probe over controls (44).This trial was much larger than the previous two and included a detailed description of bleeding points (Table 6).Two hundreJ and four patients with stigmata of recent hemorrhage were randomizcJ to bipolar treatment or control.Overall 17 of l O l ( 17%) treated patients continued to bleed or rebled within five days, versus 34 of 103 (34%) in the control group (P<0.01).The differences were most marked in groups with visible vessels, although statistical analysis was not done on subgroups.There was no significant difference in surgery or mortality.
In an open study (45), 53 patients with various upper gastrointestinal tract lesions were treated with a larger (3.4 mm) bipolar electrocoagulator.Bleeding was stopped in 52 of 53 patients and rebleeding noted in only 15%.Three comparative studies have since been carried out with variable results.In the first ( 46), 41 patients with stigmata of recent hemorrhage were randomize<l to electrocoagularion or control; six of20 treated and eight of21 control patients suffered from rcbleeding.There was no significant difference in any important parameter between the two groups.In another small study for active hemorrhage from various lesions (47), there was a significant reduction in the need for emergency surgery (14 over 57%, P=O.Ol).The rebleeding rares were not discussed and there was no significant difference in mortality.
A detailed randomized comparison of Y AG laser anJ 3.2 mm bipolar probe 10 patients with bleeding and non-bleeding visible vessels was performed in 1987 (29).Thisstm.lydiffered in two important respects from most previous trials.First, all ulcers were pretreated with an injection of 1:10,000 adrenaline.Also, a second treatment with the appropriate modality was at• tem.pted before patients were classific<l as treatment failures anJ treated surg1• cally.The results of this trial demonstrated excellent hemostanc ef.ficiency of both modalities (Table 7).
The bipolar probe appcarcJ to be

OTHER TREATMENT PROBLEMS
Some ulcers in all series arc endo-scop1cally inaccessihle and thus not amenable to any thermal method, eg, Juodenal ulcers wirh duodena l deformmes or pylonc stcnosis.Others arc m difficult locations where correct targetmg 1s not po:;s1ble.These locations are the posrenor/inferior duodenal bulb, the mfcnor post hulbar duodenum, the high gastnc lesser curve and the post- Rebleeding following initially successful hemostasts by a thermal method occurs m 10 to 30% of ulcers demonstrating visible vessels.A study previously discussed (29) demomtrated that the majority of these rebleec.lscan be terminated by a further endoscopic treatment, and recent experience has confinned this.

LA TE OUTCOME OF PATIENTS TREATED FOR BLEEDING ULCERS
A recent study from the authors' unit ( 51) looked at long term follow-up of patients with peptic ulcer hemorrhage treated between 1984 and 1986 as part of a randomized controlled trial of endoscopic YAG laser and heater probe.
Details on 9 3 of 131 patients were obtained for an average follow-up period of three years.A total of six patients suffered further peptic ulcer hemorrhage during this period: two of 25 bleeds ocri.;urred in the laser-treated group; three of 29 in the heater probetreated group; and one of 1 7 in the control group who did not have surgery.No patients who had surgery rcbled.Thi~ low rate of recurrent bleeding occurred despite a surprisingly high mcidence of dyspeptic symptoms -35% in patients treated with endoscopic thermal methods; 70% of these patients were taking Hz receptor antagonists.It is concluded that endoscopic therapy for peptic ulcer hemorrhage followed by long term maintenance Hz receptor antagonists provides long term hcmostasis in almost all patients.

DISCUSSION
The discovery of the visible vessel as the important risk factor for further bleeding in peptic ulcer hemorrhage has allowed the identification of patients likely to benefit from endoscopic therapy.Most studies have been performed with the YAG laser and evidence for the efficacy of this moJality ts convincing.Directcomparisonofargon and Y AG lasers from studies with identical protocols from this unit confirmed the clinical impression that the Y AG laser has superior hemostatic ability.There are less data for the other thermal modalities.Studies with heater probes have produced variable results.The authors' present study showed a trend towar<ls reduced rebleeding in patients treated with heater probe over controls, but this was not significant, whereas there was a significant reduction in rebleeding in patients treated with YAG.It is likely that type 2 error (not enough patients) was responsible for the lack of significant difference between the heater probe and control groups.The clinical impression was that heater probe was effective but not as good as laser, at least in the authors' hands.
The other two prospective stuJies with the heater probe which characterized the bleeding point carefully both confirmed efficacy of this device.The results of one study (38) were particularly encouraging.The small number of tnals conducted with bipolar and monopolar electrocoagulation have shown both of these modalities to be better than control, but few data are available on comparison with the YAG laser.The only comparative study (29) was well designed and gave excellent results for both Y AG laser and bipolar; both groups were preinjected with adrenaline.Subsequent studies (52,53) have shown that adrenaline mjectinn alone 1s effective in reducing furrher bleeding from peptic ulcers, although the combinatton of adrenaline and a thermal method achieved better results.
Results with this regimen 111 the second study were comparable with a combmation of adrenaline and Y AG l~cr.If these results are confirmed by other centres, excellent hemostasis ts possible without a thermal modality and could therefore be used m nonspecialist centres with considernble benefit.It 1s likely that all of the techniques described in this paper are effective at least to some extent.The evidence is most convincing for the NdYAG laser, bur it is unlikely that any major dif• ference between the modalities will emerge.The skill and perseverance of the endoscopist is almost certainly more important than the particular modality employed.Personal preference and ex• periencc may dctennine which method is most appropriate for each endoscopist.The authors' preference 1s for injection with adrenaline followed by NdY AG laser therapy.Of the last 21 patients m this hospital with peptic ulcer hemorrhage and visible vesseb treated with this combination, only three suffere<l further bleeding.Per• manent hcmostasis wa~ achieved in all three patients after repeat endoscopic therapy.These results are comparable with those achieved by the group who pioneered the combination of inJec• tion and a thermal modality (29) CAN J GASTROENTEROL VOL 4 No 9 DEC EMAER 1990 Endoscopic thermal treatment 3. Dmnficld MW.Medical or surgical crcatment ofhaemaremes,s and mclaena.JR Coll Phys Lon 1979; 1 3 :84-6.4. Allan R, Dykes P. A study of the factors mfluencing mortality rates from g,1stroinresrinal hemorrhage.QJ Med 1976; 180:533-50. 5. Avery Jone, F. Haematcmesis and mclaena.BrMedJ 1947;2:441.6. Hunt PS.The management of bleeding chrornc peptic ulcer-A ten year

Table I .
In contrast ro most prcc.edingstudies, patients were entered mto this study only if laser therapy was rhoughr to he techmcally possible at endoscopy.Thus treatment failures were tru ly due to inefficacy of laser treatment anJ not endoscopic technique.Such factors were 1dent1fied in 5% of pat1enrs in this study.Mose subsequent mals excluded these difficult cases.

TABLE 2
Results for patients with visible vessels randomized to Y AG laser or no treatment (Mcleod et al 1983) (13) Rebleedlng P<(J.02.Emergency surgery P<0.02.Morto//ty no significant difference 38 pattents tn this group trl'ateJ with laser, but in only 25 of 32 or 77°/o of wnrrnb.)Rehleedtngoccurred in two of 38 treated pat 1ents and five of 25 controls (not sign1f1cant).Emergency surgery was required111four control patients and one patien t treated with laser (not s1gn1ficant).A separate group of patients with peptic ulcers with stigmata of recent hemorrhage ( visible vessel or fresh clot) were randomized to laser or no treatment: three of 14 rehled in the laser group and seven of22 m the control; this aga in did nor reach statis-A reduction m rebleeding ma small group of patients with visible vesseb treated with Y AG was first reported 111 1983 (13).All pattems had sigm.•ofmaim blood loss (shock or hemoglobin

TABLE 3
Results of NdY AG laser in peptic ulcers with stigmata of recent hemorrhage n Number of patients: SRH Stigmata of recent hemorrhage; W Visible vessels, NS Not significant

TABLE 6
Patients with peptic ulcers treated with 2.3 mm bipolar probe or no treatment (O'Brien et al 1986) (44) Rebleedingwas significantly less common in the treated group (tine of 16) than in controls ( 11 of2 1 ).There was no difference in morcaltry.A simi lar Hudy (50) in 3 l , and 1t will clearly be difficult to improve on them.