Experimental Comparison of Endoscopic Yttrium-Aluminum-Garnet Laser, Electrosurgery, and Heater Probe for Canine Gut Arterial Coagulation

Transcription

1 GASTROENTEROLOGY 1987;92:111-8 ALIMENTARY TRACT Experimental Comparison of Endoscopic Yttrium-Aluminum-Garnet Laser, Electrosurgery, and Heater Probe for Canine Gut Arterial Coagulation Importance of Compression and Avoidance of Erosion JAMES H. JOHNSTON, DENNIS M. JENSEN, and DAVID AUTH University of Mississippi Medical Center, Jackson, Mississippi; University of California Los Angeles Center for Health Sciences, and Wadsworth Veterans Administration Medical Center, Los Angeles, California; and Squibb Medical Products, Bellevue, Washington, and University of Washington, Seattle, Washington We compared five endoscopic thermal devices that have been used clinically for control of arterial bleeding in the gastrointestinal tract: neodymium:yttrium-a1uminum-garnet [YAG) laser, electrofulguration, monopo1ar and bipolar electrocoagulation, and heater probe. In canine models we determined the ability to coagulate arteries of varying size and depth. The most effective method for coagulation of inedium size arteries was first to occlude the vessel by compression, then to apply heat to seal it [coaptive coagulation). For the contact probes [monopo1ar, bipolar, and heater probes), the depth of tissue coagulation was controlled by varying probe appositional force and energy, and coagulation of deep arteries was possible. Undesirable erosion of tissue and vessels was noted with electrical sparking from the monopolqr electrode and with Y AG laser at high power density. In contrast, heater probe and bipolar electrocoagulation did not pro- Received May 23, Accepted October 2, Address requests for reprints to: Dennis M. Jensen, M.D., CHS, Gastrointestinal Division, UCLA Center for the Health Sciences, Los Angeles, California 924. These studies were performed at the Center for Ulcer Research and Education, Wadsworth Veterans Administration Medical Center, Los Angeles, California, and at the University of Mississippi Medical Center, Jackson, Mississippi. Support included the Veterans Administration Medical Research Service and National Institutes of Health grant AM to the Center for Ulcer Research and Education. The authors thank Janet Elashoff for statistical assistance; Mary Crump, Sandra Kasko, and Virginia Kennedy for secretarial assistance; and Cooper Lasersonics for the use of a neodymium:yttrium-aluminum-garnet laser by the American Gastroenterological Association /87/$3.5 duce tissue erosion at any instrument setting. In this comparative study of arterial coagulation, overall ranking was as follows: heater probe = bipolar electrocoagulation > monopo1ar electrocoagulation > neodymium:yag laser > e1ectrofu1guration > control. In our opinion these data should assist the clinician or investigator who plans to coagulate arterial bleeding lesions such as peptic ulcers, although further clinical research will be needed to verify these experimental results. A number of thermal modalities have emerged as promising candidates for effective endoscopic coagulation of gastrointestinal bleeding. These methods include neodymium:yttrium-aluminum-garnet (Y AG) laser, monopolar (MP) and bipolar (BP) electrocoagulation, electro fulguration, and heater probe (HP) (1-11). Previous experimental comparisons of these methods have focused on efficacy, using an acute bleeding ulcer model (12), and relative safety as assessed by limited depth of tissue injury (13-19). However, this ulcer model has little resemblance to major arterial bleeding from peptic ulcers observed clinically. Several clinicians have emphasized that the endoscopic finding of a protruding "visible vessel" in a peptic ulcer base is frequently associated with further major hemorrhage (2-22). Pathologic material from gastric ulcers has shown that the "visible vessel," seen endoscopically as a raised red or whitish mound in a peptic ulcer base, in the Abbreviations used in this paper: BP, bipolar; HP, heater probe; MP, monopolar; YAG, yttrium-aluminum-garnet.

2 112 JOHNSTON ET AL. GASTROENTEROLOGY Vol. 92, No.5, Part 1 majority of cases is actually a sentinel clot extending upward from a disruption in the wall of an underlying "invisible" artery (23,24). The mean diameter of the underlying artery was.7 mm (range mm) and the location of the bleeding artery was submucosal in 19 of 29 patients and serosal in 1 patients (23). Considerable clinical experience has been reported for endoscopic hemostasis with MP electrocoagulation and lasers (1,3,6,7,25,26). There has been less clinical experience with BP electrocoagulation and HP (27-34). Studies comparing these devices for arterial coagulation are needed. Certain important factors, such as vessel size, depth, and blood flows are most difficult to evaluate clinically and require careful experimental study. The purposes of the present study were to compare five endoscopic instruments regarding the ability to directly coagulate an exposed artery and the ability to coagulate an artery located deeper within the tissue. Materials and Methods Treatment Modalities YttTium-aluminum-garnet laset. A 12-W YAG laser (model 8, Cooper Lasersonics, Sunnyvale, Calif.) with 6-,um quartz lightguide and coaxial gas jet was employed. The full angle of divergence was 9.4. The treatment distance was controlled by an adjustable stand. Coaxial gas jet flow was 2 mlls. Electrofulguration. Electrofulguration was produced with a modified needle electrode (.7-mm diameter) and Valleylab SSE2-K electrosurgical unit (Valleylab, Boulder, Colo.) in the pure coagulation mode. There was frequent water irrigation by a hand-held syringe to keep the target field wet. The treatment distance was adjusted to produce a sparking effect. Monopolar electrocoagulation. A monopolar probe (model CD-3L, Olympus Corp., Lake Success, N.Y.) of 2.2-mm OD with 1.2-mm central irrigation channel was used with a Valleylab SSE2-K electrosurgical unit in the pure coagulation mode. Intermittent water irrigation was used to keep the target field wet. Appositional force with the contact probes was controlled by a hand-held force gauge (4). Bipolar electrocoagulation. The 25-W BICAP unit (American ACMI, Stamford, Conn.) with large probe (3.3-mm OD with distal water irrigation through a central channel) was employed (8). Power, pulse duration, and appositional force were controlled. Heater probe. A prototype heater probe unit (now manufactured by Olympus Corp.) with large endoscopic probe (3.2-mm OD, rounded tip with teflon coating, and water irrigation ports 1 cm from the probe tip) was employed (9). Energy and appositional force were controlled. A typical HP application of 3 J lasted about 8 s. Table 1. Instrumental Settings for the Experimental Studies "Usual" settings High settings Yttrium-aluminum-garnet laser Power (W) 75 1 Spot diameter (mm) Power density (W/cm) ,738 Pulse duration (s) 1 1 Fulguration Power (coagulation setting) 5 1 Pulse duration (s) 1 1 Monopolar probe Power (coagulation setting) 3 1 Pulse duration (s).5.5 Appositional force (g) 1 1 Bipolar probe Power setting 1 1 Pulse duration (8) 2 2 Appositional force (g) 1 1 Heater probe Energy U) 3 3 Appositional force (g) 1 1 In experimental studies of blood vessel coagulation, two treatment settings were tested for the five thermal devices. "Usual" settings for each device were chosen to represent settings typically used in clinical treatment of gastrointestinal bleeding. High power density for each particular device was tested also. Canine Model Sixteen healthy fasted adult mongrel dogs were used for these acute experiments. The anesthesia protocol was acepromazine maleate and atropine sulfate intramuscularly followed by intravenous sodium pentobarbital. Animals received endotracheal intubation, and blood loss was replaced with intravenous normal saline. Laparotomy was performed with standard surgical techniques. Experiment 1: direct arterial coagulation. The ability of each device to directly coagulate an exposed artery was assessed. Treatment was directed to a single spot on each artery. "Usual" instrumental settings (similar to those used clinically for coagulation of bleeding ulcers) were employed for each device (Table 1), with the exception of MP electrocoagulation, which was applied with short pulses of <.5 s to minimize electrical sparking. In section E of this experiment, high power densities were also evaluated. Contact probes were applied with sufficient force (1-3 g) to occlude the target artery, unless conditions are otherwise specified. Repeated treatment pulses were delivered until coagulation was obtained, or a maximum of 1 pulses. After treatment, the artery was severed at the treatment site with a scalpel. Coagulation was considered successful only if there was no bleeding from the cut artery. For controls an adjacent artery of each size being evaluated ( mm) had apposition force applied for 3 min with a probe (HP or BP-parts A-C and E) or a glass slide (part E) without application of heat. After 3 min, the artery was cut with a scalpel (for parts A and C-E) and observed for 3 min. Ligation was performed on arteries

3 May 1987 ENDOSCOPIC COAGULATION OF ARTERIES 113 Table 2. Percentage Efficacy of Direct Coagulation of Arteries C. Probe A. Intact arteries of varying size without D. Laser E. High (diameter in mm) compression of with glass power B. Bleeding bleeding slide density artery artery compression settings YAG laser 1 42 a 2 a a NA 1 Ob Fulguration loob Ob Ob a Ob NA NA Ob Monopolar 1 loob 87 b 67 b 8 b Ob NA Ob Bipolar a NA 1 Heater probe a NA 1 Control a a a a a a a a NA, not applicable. YAG, yttrium-aluminum-garnet. Fifteen arteries were treated per entry. Numbers in the chart represent the percentage of arteries successfully coagulated. Contact probes (MP, BP, HP) were applied with appositional force sufficientto occlude the artery (1-2 g) in A, B, and E; light touch «1 g) was used in C. For controls, appositional force was applied with a probe (HP or BP) or glass slide for 3 min without application of heat. Laser was applied as a noncontact device in A, B, C, and E; ancillary glass slide compression of the target artery was added to laser treatment in D. "Usual" instrumental settings of each device were used in A-D except with MP electrocoagulation, for which very short «.5 s) pulses were used. High power density settings were employed in E (see Table 1). One-millimeter mesenteric arteries were treated in B-E. Statistical significance: In A and B, YAG and fulguration were significantly less effective than MP and BP electrocoagulation and HP except for.25-mm arteries. In E, YAG, fulguration, and MP electrocoagulation were significantly less effective than BP electrocoagulation and HP (p <.5). a Thirty-six arteries treated in this particular entry. b Erosion of vessel with induced bleeding. (control and treated) that cqntinued to bleed after the observation period. The following factors were stmlied: A. Arterial size. The outside di<\ffieter of intact mesenteric and serosal arteries was measufl;ld with a micrometer. Intact arteries of.25-,.5-, 1.-, and 1.5-mm diameter were treated with each device. For bipolar electrocoagulation and heater probe, 2.-, 2.5-, and 3.-mm arteries were also coagulated. B. Active bleeding. To assess efficacy in treating a bleeding artery, active bleeding was produced by puncturing a l-mm mesenteric artery with an 18-gauge needle. Treatment with each device was delivered directly to the bleeding site. Treatment was judged successful only if active bleeding was halted with 1 or fewer treatment pulses. C. Contact probe treatment without vessel compression. Our usual technique in coagulating arteries with the contact probes (MP, BP, HP) involved pressing against the artery with sufficient force to occlude it prior to heat delivery. In this part of the experiment, the contact probes were applied with only light touch «1 g of appositional force) to intact l-mm mesenteric arteries. D. Laser treatment with ancillary vessel compression. Laser treatment is usually performed without mechanical contact with the vessel. In this part of the experiment, ancillary vessel compression was added to laser treatment by pressing a clear glass slide against the artery to occlude it, then directing the laser light through the transparent glass slide to heat the compressed artery. E. Effect of high power density. This part of the experiment assessed the effect of treatment of an intact l-mm mesenteric artery with each device using high power density settings (see Table 1). Experiment 2: coagulation of a deep underlying artery. This experiment tested the ability of each device to coagulate an artery located several millimeters beneath the surface. For this study, treatment was applied to the mucosal side of opened intact small intestine in an attempt to coagulate an underlying serosal artery. The serosal artery had a mean outside diameter of.25 mm and was located -3 mm below the mucosa. For adjacent control vessels, appositional force with heater probe was applied (light, moderate, or heavy) for 3 min but no heat was applied. "Usual" instrumental settings were employed with each device (Table 1) with the exception of MP electrocoagulation, which was employed with coagulation power settings of 7- and l-s pulses. With the contact probes, three different appositional forces were tested: (a) light touch «1 g), no tamponade of the deep serosal artery or its paired vein; (b) moderate force (15-2 g), enough to tamponade the accompanying vein but not the serosal artery; (c) heavy force (1-12 g), sufficient to tamponade the paired serosal artery and vein. Successful coagulation was assessed by puncture of the target artery with a needle. Statistical Methods Differences in efficacy between coagulation methods were tested using log linear models with dog as a factor in the statistical program BMDP4F (35). Results Experiment 1 None of the control arteries of any size ( mm) stopped bleeding spontaneously and all were ligated after 3 min of observation (see Table 2). Yttrium-aluminum-garnet laser was effective for coagulation of small (.25 mm) arteries but was not effective for larger vessels. Yttrium-aluminum-garnet laser heating resulted in arterial shrinkage with

4 114 JOHNSTON ET AL. GASTROENTEROLOGY Vol. 92, No.5, Part 1 coagulation of only small arteries. The importance of eliminating the arterial "heat sink" by vessel compression was demonstrated dramatically by the ease of coagulation of medium size arteries with the laser when glass slide tamponade was added (part D). Yttrium-alum inurn-garnet laser induced arterial spurting by erosion of the vessel wall when high power densities were employed (part E). Electrofulguration was poor in arterial coagulation and was only effective with arteries of.25-mm diameter. Electrical sparking frequently caused vessel erosion with induced bleeding. Heater probe and BP and MP electrocoagulation were significantly more effective than Y AG and fulguration in coagulation of arteries of mm diameter (part A). To achieve efficacy with these contact probes, complete occlusion of the vessel by probe apposition was essential. Attempts to coagulate an artery with light touch of the probes were consistently unsuccessful (part C). The schematic representation of these results is given in Figure t. The MP probe was limited by vessel erosion produced by electrical sparking, despite our considerable attempts to avoid sparking. This erosion effect interfered with arterial coagulation of arteries larger than.5 mm in diameter. In contrast, no erosion of vessels or induced bleeding was observed with BP electrocoagulation or HP even with unlimited pulses or continuous application. Heater probe and BP electrocoagulation were similar in efficacy for coagulation of arteries of different sizes (Figure 2 and Table 2). All arteries <1.5 mm in diameter were easily coagulated with either probe. Most (93% BP, 8% HP) 2-mm arteries could be effectively coagulated, whereas 6% of 2.5-mm arteries could be coagulated, and only 13% of 3.-mm were successfully coagulated with either HP or BICAP. Overall, the order of effectiyeness in direct coagulation of exposed arteries was HP = BP > MP > Y AG > fulguration> control.. Experiment 2 Monopolar and BP electrocoagulation and HP were each significantly more effective than Y AG and fulguration in coagulation of a deep underlying artery (see Table 3). None of the controls resulted in clotting of the transmural artery. All attempts to transmurally coagulate small serosal arteries with Y AG and fulguration were unsuccessful. Deep erosion of the small intestinal wall occurred before the underlying serosal vessel could be coagulated with these noncontact modalities. Coagulation with HP and BP and MP electrocoagulation could be achieved only with complete arte- rial tamponade. With appositional pressure sufficient to tamponade the companion vein but not the artery, HP and BP and MP electrocoagulation were each ineffective. Transmural tamponade and coagulation of serosal arteries with HP was typically achieved with one or two pulses, whereas BP and MP electrocoagulation each usually required two to four pulses for coagulation. Higher than "usual" MP power settings were required for efficacy (Table 1). Discussion This study compared the hemostatic capabilities of five endoscopic thermal devices with control. Tissue heating may be produced by electrical current passing through resistive tissue, by tissue absorption of laser light energy, or by diffusion of heat from another source. Tissue heated to 6 C turns white due to protein coagulation. However, thermal coagulation of solid tissue protein should not be confused with ability to "coagulate" a bleeding artery. The importance of the heat sink effect of flowing arterial blood has been underestimated both experimentally and clinically in modern gastroenterology. To illustrate, with a typical 1-mm artery that has blood flow of 3 mllmin, each red blood cell travels at a mean velocity >2 ft/s. During laser treatment, each red cell is barely warmed as it passes through the laser beam (calculated temperature rise of only 4 C). Thus, thermal energy is effectively dispersed by the rapid flow of arterial blood. With the laser, the most important factor in arterial coagulation is vessel shrinkage. Temperature rise to 8-9 C causes contraction of collagen in the blood vessel wall, which results in vessel shrinkage and reduction of blood flow (36). We found that laser treatment was only effective for direct coagulation of.25-mm exposed arteries, and became progressively less effective with larger arteries (Table 2). We did not confirm earlier observations that the YAG laser coagulates larger arteries (37). A powerful ancillary effect is added when a contact probe is used to compress the target artery before heat delivery. Vessel compression eliminates the heat sink effect; coaptation of vessel walls also facilitates sealing by heat. We found that when HP and BP electrocoagulation were used with sufficient appositional force to tamponade blood flow, they were consistently effective in direct coagulation of mm arteries (Table 2). Arteries larger than 2. mm could not be consistently coagulated even with firm compression and high power settings. In contrast, when these probes were employed with only light touch, they were ineffective for this size of arteries. Similarly, the addition of vessel compression by a glass slide to Y AG laser treatment enabled

5 May 1987 ENDOSCOPIC COAGULATION OF ARTERIES 115 A 2 3 SIIIIIII lectlon through Imlll Intlltln. Laser beam Tillue.rOllon B SUblllUCOIlllrtlry lone tlllue COllullllon ("whlt.lpol"i 2 Art.rlll "h'lt.ink" due to npld blood flow 3 V... I shrlnklge 4 Thermll probe Probe "foot print" Plllnllrtlry _pllltillu. COIIullIllII V..., compr.aslon Ind tlmponld. by probe Thermal leiling Df the compra.. d va.el ~ COIIulllld Irtery Figure 1. Schematic comparison of arterial hemostasis by zonal heating and coaptive coagulation. A. (1) Diagram of sagittal section through small intestine with submucosal artery shown. Direction of blood flow is indicated by arrow. (2) Yttrium-aluminumgarnet laser treatment with "usual" settings produces full-thickness coagulation of tissue protein (white spot), but the submucosal blood vessel is not coagulated due to heat sink effect. (3) Yttrium-aluminum-garnet laser with higher energy density settings produces shrinkage of the target vessel, but causes undesirable superficial vaporization. B. (1) Heater or bipolar probe, applied with light touch, produces a superficial zone of coagulation, but the submucosal artery is unaffected. (2) Heater or bipolar probe is applied with force suffcient to compress and occlude the underlying artery, prior to delivery of heat. (3) With vessel tamponade and coaptation, heat is delivered to seal the compressed vessel. (4) After deep treatment with heater or bipolar probe, a probe "footprint" is left in the tissue, representing compressed and coagulated tissue. coagulation of arteries up to 1.5 mm. We conclude that the most important ancillary factor for effective arterial coagulation is mechanical vessel compression, irrespective of which particular device produces the heat. The concept of vascular compression and tamponade is not new. In the 196s, Sigel and Dunn (38) and Sigel and Hatke (39) performed an elegant set of experiments contrasting obliterative coagulation (heating the vessel and surrounding tissue by light touch of a monopolar electrode to cause vessel shrinkage for hemostasis) with coaptive coagulation

6 116 JOHNSTON ET AL. GASTROENTEROLOGY Vol. 92, No.5, Part 1 1 w ~ < LL...J ::1 (!) ~< zo WO 5 a: w~ C1. a: w ~ a: <.25 ARTERY DIAMETER IN MILLIMETERS Figure 2. Histogram of the efficacy of BlCAP (open bar) and heater probe (cross-hatched bar) for coagulation of mesenteric arteries, expressed as a percentage vs. artery diameter in millimeters. (using a hemostat to tamponade blood flow and coapt the vessel walls, followed by application of electrocautery to thermally seal the vessel). They concluded that obliterative coagulation was only effective with small «1 mm) arteries, whereas coaptive coagulation was effective with arteries up to 2-4 mm in size. A second factor, which is distinctly undesirable in the setting of gastrointestinal hemorrhage, is tissue erosion (ablation or cutting of tissue). Erosion occurs by rapid boiling of tissue water (1 C), producing tiny steam pockets that explode causing disruption of the tissue (4). At much higher temperatures, dried tissue can actually be eliminated by vaporization, combustion, or carbonization. These events are possible with lasers and electrical sparking, each of which can rapidly generate tissue temperature in the several thousand degree range (41). An erosive tendency with these devices translates clinically to a potential for induced bleeding by cutting into blood vessels as well as acute perforation of the intestinal wall. Such an erosive effect may account for the precipitation of bleeding with Y AG laser treatment of nonbleeding visible vessels (42-44). In contrast, tissue dessication (drying) may occur by evaporation of water «1 C) or by slow boiling (1 C) at a rate that allows steam vapor to diffuse from tissue without disruption of structure. Each thermal device can produce tissue coagulation and dessication. Only BP electrocoagulation and HP have safeguards to prevent tissue erosion. With electrodessication (BP, MP), tissue temperature clamps at 1 C until tissue water is boiled away. Once tissue is dry, electrical flow markedly drops because of high tissue resistance. At this point, the powerful MP generator (peak voltage >5 V) can drive electrons across a superficial zone of dry tissue by electrical sparking, thus producing tissue erosion. In contrast, maximal BlCAP generator output is limited to ~5 V, which is insufficient to produce sparking. Thus, thermal changes in tissue conductivity combined with limited generator voltage prevent tissue erosion with BP. The HP avoids tissue erosion by controlling tem- Table 3. Ability to Coagulate a Deep Underlying Artery Yttrium-aluminum-garnet laser Fulguration Monopolar probe Bipolar probe Heater probe Control c Noncontact a vs. <1 g (light touch)b Probe appositional force 15-2 g (occludes vein, not artery) 1-12 g ( occludes artery) NA NA NA NA NA, not applicable. Ten arteries were treated per entry. Numbers in the chart represent the percentage of arteries successfully coagulated. a Noncontact refers to YAG and fulguration application. b Light touch was applied with the other methods. C Control was application of transmural force with heater probe for 3 min without heat.

7 May 1987 ENDOSCOPIC COAGULATION OF ARTERIES 117 perature directly (9). Maximal internal probe temperature is ~25 C. Tissue temperature clamps at 1 C until tissue water is boiled away and then may rise above 1 C. Tissue heating by thermal diffusion is slow compared with near instantaneous heating by lasers or electrical sparking. Tissue erosion was most frequent during fulguration. When using MP electrocoagulation, we were unable to avoid a small sparking tendency, which at times caused erosion of the vessel wall with induced bleeding when the probe was placed directly on a vessel. By constant perfusion of water through the MP probe tip, sparking can be prevented without affecting coagulation (similar to the electro-hydrothermo electrode) (45). However, constant water irrigation may cause undesirable intramural dissection if the probe is used for forceful tamponade. With high power density, YAG laser produced rapid and deep tissue erosion with attendant vessel wall disruption and induced bleeding. A third major requirement for successful arterial hemostasis is effective delivery of coagulative heat to the depth of the underlying artery. In the clinical setting, the bleeding artery is often at or just below the peptic ulcer base (23). At times, however, the target artery may be deeper in the tissue (24). Even though the depth of tissue heating with Y AG laser may be substantially greater than with other modalities, the effective depth for arterial coagulation was quite restricted (Table 3). We have not found that modifications of Y AG beam configuration, including testing a convergent focused beam, significantly altered the effective depth of vessel coagulation (unpublished observations). In contrast, compression of tissue with an endoscopic probe is an efficient and controllable way to increase coagulation depth. Using the technique of vessel tamponade, it was not difficult to coagulate deep lying vessels (Table 3). For vessel coagulation, there were no consistent differences in efficacy between HP and BP electrocoagulation (Figure 2). Neither device could consistently coagulate arteries larger than 2 mm. However, both devices were significantly more efficient than any other thermal device for.5-2.-mm arteries. As the mean diameter of arteries underlying clinical bleeding gastric ulcers was.7 mm in pathologic study of bleeding gastric ulcers with a range of mm (23), one might expect BP electrocoagulation and HP to be quite effective clinically. Our preliminary clinical results with HP and BP electrocoagulation indicate that endoscopic tamponade and hemostasis of bleeding peptic ulcers are feasible and effective (32-34). After thermal coagulation of peptic ulcers with any endoscopic device, one might expect that the tissue necrosis would result in a larger peptic ulcer, a delay in healing, or delayed slough and severe rebleeding. All of these effects of thermal coagulation have been reported for the small bowel or colon (44,46-48), and in our experience clinically, may be observed in the catabolic patient with a bleeding ulcer who is not expected to heal. On the contrary, most patients with bleeding peptic ~lcers treated with careful target coagulation by Y AG or argon laser (1,21-23,25,26,32,42)' electrocoagulation (6,7, 11,27-3,33), or heater probe (28,32,33) have been reported to have high rates of permanent hemostasis and subsequent ulcer healing. Treatment with thermal devices is easiest in an experimental setting-with clear field, excellent exposure, en face orientation, and precise control of treatment distance (laser) and appositional force (contact probes). These ideal conditions are difficult or impossible to achieve in the clinical endoscopic setting of active arterial bleeding from a peptic ulcer. Although there is no model that precisely matches the clinical situation, these laboratory experiments provide comparisons of different modalities with precise control of the important variables. Yttriumaluminum-garnet, fulguration, and MP electrocoagulation each had major limitations with arterial coagulation, even with these "ideal" conditions. It has been our experience that these limitations are further magnified in the clinical setting. It is our opinion that the nonerosive contact probes represent a major advance for the endoscopic treatment of gastrointestinal hemorrhage from arterial lesions. References 1. Kiefhaber P, Nath G, Moritz K. Endoscopic control of massive gastrointestinal hemorrhage by irradiation with a high power neodymium-yag laser. Prog Surg 1977;15: Johnston JH, Jensen DM, Mautner W, Elashoff J. YAG laser treatment of experimental bleeding canine gastric ulcers. Gastroenterology 198;79: Johnston J. Specific treatment techniques for massive upper gastrointestinal bleeding. In: Fleischer D, Jensen D, Btight Asare P, eds. Therapeutic laser endoscopy in gastrointestinal disease. Boston: Martinus Nijhoff, 1983: Piercey JRA, Auth DC, Silverstein FE, et al. Electrosurgical treatment of experimental bleeding canine gastric ulcers: development and testing of a computer control and a better electrode. Gastroenterology 1978;74: Dennis MB, Peoples J, Hulett DC, et al. Evaluation of electrofulguration in control of bleeding of experimental gastric ulcers. Dig Dis Sci 1979;24: Papp JP. Electrocoagulation. In: Papp JP, ed. Endoscopic control of gastrointestinal hemorrhage. Boca Raton, Fla.: CRC Press, 1981: Gaisford WE. Endoscopic electrohemostasis of active upper gastrointestinal bleeding. Am J Surg 1979;137: Auth DC, Gilbert DA, Opie E, Silverstein FE. The multipolar probe--a new endoscopic technique to control gastrointestinal bleeding. Gastrointest Endosc 198;26:A Protell RL, Rubin CE, Auth DC, et al. The heater probe: a new

8 118 JOHNSTON ET A1. GASTROENTEROLOGY Vol. 92, No.5, Part 1 endoscopic method for stopping massive gastrointestinal bleeding. Gastroenterology 1978;74: Gilbert DA, Silverstein FE, Auth DC, Rubin CE. Nonsurgical management of acute nonvariceal upper gastrointestinal bleeding. In: Spaet TH, ed. Progress in hemostasis and thrombosis. Volume.4. Grune & Stratton, 1978: Jensen D. Endoscopic control of gastrointestinal bleeding with non-laser devices. In: Fleischer D, Jensen D, Bright Asare P, eds. Therapeutic laser endoscopy in gastrointestinal disease. Boston: Martinus Nijhoff, 1983: Protell RI., Silverstein FE, Piercey J, Dennis M, Spake W, Rubin CE. A reproducible animal model of acute bleeding ulcer-the "ulcer maker." Gastroenterology 19713;71: Johnston JH, Jensen DM, Mautner W. Comparison of endoscopic electrocoagulation and laser photocoagulation of bleeding canine gastric ulcers. Gastroenterology 1982; 82: Silverstein FE, Protell RL, Gilbert DA, et a1. Argon vs. Nd Y AG laser photocoagulation of experimental canine gastric ulcer. Gastroenterology 1979;77: Machicado GA, Jensen DM, Tapia JI, Mautner W. Treatment of bleeding canine duodenal and esophageal ulcers with argon laser and bipolar electrocoagulation. Gastroenterology 1981;81: Jensen DM, Machicado GA, Tapia I, Mautner W. Comparison of argon laser photocoagulation and bipolar electrocoagulation for endoscopic hemostasis in the canine colon. Gastroenterology 1982;83: Jensen DM, Tapia JI, Machicado GA, Beilin DB, Silpa M. Comparison of electroco~gulation and heater probe for hemostasis in the canine colon. Gastrointest Endosc 1982;28:A Swain CP, Mills TN, Dark JM, et a1. Comparative study of the safety and efficacy of liquid and dry monopolar electrodes in experimental canine bleeding ulcers using computerized energy monitoring. Gastroenterology 1984;86: Swain CP, Mills TN, Shemesh, et a1. Which electrode? A consumer's guide to endoscopic electrocoagulation of upper gastrointestinal bleeding. Gut 1983;24:A Griffiths WI, Neumann DA, Welsh JD. The visible vessel as an indicator of uncontrolled or recurrent gastrointestinal hemorrhage. N Engl J Med 1979;3: Storey DW, Bown SG, Swain CP, et a1. Endoscopic prediction of recurrent bleeding in peptic ulcers. N Engl J Med 1981;35: Vallon AG, Cotton PB, Laurence BH, et a1. Randomized trial of endoscopic argon laser photocoagulation in bleeding peptic ulcers. Gut 1981;22: Swain CP, Storey DW, Bown SG, et a1. Nature of the bleeding vessel in recurrently bleeding gastric ulcers. Gastroenterology 1986;9: Johnston JH. The sentinel clot and invisible vessel: pathologic anatomy of bleeding peptic ulcer. Gastrointest Endosc 1984;3: Rutgeerts P, Geoboes K, Vantrappen G. European experience with Nd:YAG and argon laser for therapy of upper gastrointestin;;j.i bleeding. In: Fleischer D, Jensen D, Bright-Asare P, eds. Therapeutic laser endoscopy in gastrointestinal disease. Boston: Martinus Nijhoff, 1983: Bown SG. Controlled studies of laser therapy for hemorrhage from peptic ulcers. Acta Endoscopica 1985;15: Gilbert DA, Verhoeven T, Jessen K, Bown SG, Bowers JH, Papp JP. A multicenter clinical trial of the BlCAP probe for upper gastrointestinal bleeding. Gastrointest Endosc 1982; 28:A Jensen DM, Machicado GA, Tapia JI, Beilin DB. Clinical hemostasis with heater probe or bipolar electrocoagulation for severe gastrointestinal bleeding. Gastrointest Endosc 1983;29: Winkler WP, Comer G, McCray RS. Initial experience with BlCAP multipolar electrocautery in the control of upper gastrointestinal hemorrhage. Gastrointest Endosc 1983; 29:A Morris DL, Hawker PC, Keighley MRB, Dykes PW. Bipolar endoscopic electrocoagulation. Gut 1982;23:A Storey DW. Endoscopic control of peptic ulcer hemorrhage using the "heater probe." Gut 1983;24:A Johnston I, Sones J, Long B. Heater probe is superior to YAG laser in clinical endoscopic treatment of major bleeding from peptic ulcers. Gastrointest Endosc 1984;31: Jensen DM, Machicado GA, Silpa ML. Argon laser vs. heater probe or BlCAP for control of severe ulcer bleeding. Gastrointest Endosc 1984;3: Jensen DM, Machicado GA. BlCAP vs. heater probe for hemostasis of severe ulcer bleeding. Gastrointest Endosc 1986;32: Dixon WI, ed. BMDP statistical software. Berkley, Calif.: University of California Press, Gorisch W, Boergen KP. Heat-induced contraction of blood vessels. Lasers Surg Med 1982;2: Dwyer R. Treatment techniques for massive upper gastrointestinal bleeding, general considerations. In: Fleischer D, Jensen D, Bright-Asare P, eds. Therapeutic laser endoscopy in gastrointestinal disease. Boston: Martinus Nijhoff, 1983: Sigel B, Dunn MR. The mechanism of blood vessel closure by high frequency electrocoagulation. Surg Gynecol Obstet 1965;121: Sigel B, Hatke F1. Physical factors in electrocoaptation of blood vessels. Arch Surg 1967;95: Hall RR, Beach AD, Baker E, et a1. Incision of tissue by CO 2 laser. Nature 1971;232: O'Shea DC. Callen WR. Rhodes WT. Application of laser power. In: O'Shea DC, ed. An introduction to lasers and their applications. Reading, Mass.: Addison Wesley, Rutgeerts P, Vantrappen G, Broeckhaert L. et a1. Controlled trial of YAG laser treatment of upper digestive hemorrhage. Gastroenterology 1982:83; Krejs GJ, Little KH, Westergaard H, et a1. Laser photocoagulation for treatment of acute peptic ulcer bleeding: a randomized controlled clinical trial (abstr). Gastroenterology 1985; 88: Johnston J. Complications of endoscopic laser therapy. In: Fleischer D, Jensen D, Bright-Asare p. eds. Therapeutic laser endoscopy in gastrointestirial disease. Boston: Martinus Nijhoff.1983: Matek W, Fruhmorgen WM, Kaduk B. Reidenbach H. Bodem F, Demling 1. Modified electrocoagulation and its possibilities in the control of gastrointestinal bleeding. Endoscopy 1979;4: Jensen DM, Tapia JI, Machicado GA, Beilin DB, Silpa M. Comparison of electrocoagulation and heater probe for hemostasis in the canine colon. Gastrointest Endosc 1982;28: Jensen DM, Tapia JI, Machicado GA, Beilin DB. Hydrothermal probe. hot biopsy forceps and YAG laser for hemostasis in the canine colon. Gastrointest Endosc 1983;29: Jensen DM. Machicado GA. Bleeding colonic angioma: endoscopic coagulation and follow-up (abstr). Gastroenterology 1985;88:1433.