Secondary Logo

Journal Logo


Therapeutic Intervention for Nonvariceal Gastrointestinal Hemorrhage

Wyllie, Robert; Kay, Marsha H.

Author Information
Journal of Pediatric Gastroenterology & Nutrition: February 1996 - Volume 22 - Issue 2 - p 123-133
  • Free

The evaluation and management of acute gastrointestinal bleeding in children are a frequent reason for emergency consultation by pediatric gastroenterologists. Endoscopic evaluation is the most rapid and accurate method of identifying the cause of acute bleeding in the majority of lesions presenting in the pediatric age group. When lesions are actively bleeding or in situations where bleeding is likely, hemostasis can be achieved by a variety of endoscopic techniques. Familiarity with the various techniques is essential for the pediatric endoscopist. This review focuses on the endoscopic management of nonvariceal bleeding in infants and children.


Not every child with a gastrointestinal hemorrhage needs endoscopy. Those patients with a history compatible with an acute self-limited illness who are hemodynamically stable and have a normal hematocrit can be monitored for further evidence of bleeding. Patients who have experienced a major gastrointestinal bleed or those who rebleed usually require endoscopy as part of their evaluation. Radiologic studies such as an upper gastrointestinal series with small bowel follow-through or barium enema have a significantly decreased ability to identify the source of bleeding when compared to endoscopic evaluation (1). Other studies such as selective angiography rely upon a persistent rapid rate of bleeding to identify the origin of the bleeding. When endoscopy is performed by a well-trained endoscopist it is the most sensitive and specific technique for identifying most causes of gastrointestinal blood loss. Endoscopy in the bleeding patient requires a greater level of skill and is technically more demanding than endoscopy performed for other indications. When possible, the endoscopist should have the necessary knowledge and experience to identify the origin of the bleeding and undertake therapeutic intervention. A pediatric surgeon or colorectal surgeon should evaluate patients before emergency endoscopy for a major gastrointestinal hemorrhage. In actively bleeding patients it may be useful for the surgeon to observe the procedure. Surgical intervention should be immediately available when attempting therapeutic maneuvers in bleeding posterior bulb ulcers or other high-risk lesions.

The timing of endoscopy is determined by the severity of the bleeding. Patients who have active hemorrhage and hemodynamic instability and require transfusions need urgent endoscopy after resuscitation. In the unusual child with torrential bleeding and uncontrollable hypovolemia, endoscopy may be performed immediately prior to surgical intervention. Localization of a bleeding site to a specific anatomic area or the identification of varices may change surgical strategies. Children who have evidence of an ongoing transfusion requirement after initial attempts at resuscitation need emergent evaluation, although they are at a higher risk for complications during the endoscopic procedure. Children who have previously bled are also at an increased risk for a second bleeding episode and require more urgent evaluation (2). The timing of the endoscopy and the sedative risks must be balanced against the diagnostic and therapeutic benefits that are anticipated in the bleeding child.

Lesions that may be amenable to endoscopic therapy include (a) an ulcer with evidence of active bleeding, (b) an ulcer with oozing from beneath an overlying clot (sentinel clot), (c) an ulcer with a nonbleeding visible vessel at its base that may appear as a red, blue, or white plug or mound (sometimes referred to as a pigmented protuberance), and (d) vascular malformations (3,4). In adult studies 50% of these lesions will rebleed after an initial episode, compared to an incidence of ≤10% with other lesions including ulcers with an overlying clot without oozing or those with flat darkly pigmented red or black spots (3,5). Gastroduodenal vascular malformations such as arteriovenous malformations and telangiectasias are rare but have a high risk of recurrent bleeding. Deep ulcers on the lesser curvature of the stomach or along the posterior-inferior aspect of the duodenal bulb may be particularly at risk for severe bleeding because of their proximity to large blood vessels (6). Diffuse mucosal bleeding from duodenitis or gastritis is usually not responsive to endoscopic intervention. Colonic lesions treatable with endoscopic therapy include bleeding ulcers, angiomata, polyps and bleeding polyp stalks (7). An adherent clot in an ulcerated area of colonic mucosa with fresh blood nearby may also be a lesion at a high risk for rebleeding (8). Lesions identified at endoscopy with a high risk for rebleeding should usually be treated.


Emergent gastrointestinal endoscopy carries an increased risk of complications as compared to routine endoscopy (4). This includes a risk of aspiration of gastric contents in a stomach full of blood and a higher risk associated with sedating a potentially unstable patient with decompensated cardiopulmonary or hepatic function (4). The increased risks associated with emergent gastrointestinal endoscopy necessitate close monitoring of the airway and cautious patient sedation. In some situations, general anesthesia offers advantages to the endoscopist by providing a controlled airway, close monitoring of cardiorespiratory function, and an immobile patient. The safety and efficacy of interventional endoscopy depend on the ability of the child to cooperate during the procedure. In children who have the ability to maintain their position during endoscopy, conscious sedation is effective. After interviewing the patient and explaining the procedure, a judgment is made about whether the procedure should be performed under general anesthesia or if conscious sedation should be attempted. When conscious sedation is used, the medication dosages should be kept to a minimum to avoid impairment of the patient's cardiopulmonary status; continuous monitoring of pulse oximetry and blood pressure as well as intravenous access should be maintained. In a pediatric intensive care unit, a pediatric intensivist may assist with the sedation and monitoring. Patients undergoing therapeutic endoscopy should be volume resuscitated prior to the endoscopy if possible, in an attempt to ensure hemodynamic stability during the procedure.

Preparation for emergent endoscopy differs from that for routine procedures. The patient has not necessarily fasted prior to the procedure and has not undergone an adequate colonic cleansing in the case of colonoscopy. Gastric lavage has traditionally been used for evacuating blood from the stomach and for localizing bleeding to the upper gastrointestinal tract. Nasogastric tubes usually fail to remove all the clots and a clear nasogastric aspirate does not exclude a lesion in the upper gastrointestinal tract. Nasogastric tubes may also produce confusing mucosal artifacts (6,9). For these reasons we do not routinely perform a nasogastric aspiration in children with a history of vomiting blood who have other evidence of a significant gastrointestinal hemorrhage. In the majority of patients admitted to the intensive care unit with melena or hematochezia, we also perform a quick upper endoscopy instead of passing a nasogastric tube. Identification of a bleeding lesion in the upper tract will eliminate the delay, difficulty, and risk associated with emergent colonoscopy. In children with a history of hematemesis or coffee ground emesis without significant blood loss, gastric lavage may be useful in determining the extent of ongoing blood loss and need for endoscopy. If gastric lavage is performed, lukewarm tap water should be used instead of iced saline, since the latter has been associated with hypothermia, aspiration pneumonia, and prolongation of the bleeding time.

For colonoscopic procedures, blood in the digestive tract tends to act as a cathartic, and a full preparation may not be required. In unstable patients requiring emergent colonoscopy, attempts to administer cathartics or balanced electrolyte solutions delay the procedure and are unlikely to be helpful in clearing the colon of blood. In the hemodynamically stable patient, we administer a balanced electrolyte solution such as Golytely (Braintree Laboratories Inc., Braintree, MA, U.S.A.), Nulytely (Braintree Laboratories), or Colyte (Reed & Carnick, Jersey City, NJ, U.S.A.) to enhance our ability to visualize a single lesion that may be acting as the bleeding source. There is a risk of combustion in the colon when electrocautery is performed in patients with large amounts of fecal debris (10).


There are three basic techniques of therapeutic endoscopy for acute gastrointestinal bleeding: injection, thermocoagulation, and laser therapy. Results of comparative trials of injection, coagulation, and heater probe therapy vary according to the method used. When properly applied, each modality can achieve high rates of both initial and permanent hemostasis, generally in the range of 70-90%, and significantly decrease the need for emergency surgery (7,11-17). This has been demonstrated both in trials comparing one hemostatic agent or thermal agent against another and in trials comparing a single therapy against an untreated control group. There have been several randomized blinded trials comparing treatment groups to sham controls (5,13,18). The specific technique employed depends on the equipment availability and experience of the endoscopist.

Therapeutic endoscopy is most easily accomplished using a two-channeled therapeutic scope so that therapy (injection, coagulation, etc.) may be accomplished via one channel and simultaneous suction or irrigation performed using the second channel to keep the field in view. Unfortunately, therapeutic endoscopes are of a large diameter compared to standard endoscopes and cannot be routinely used in pediatric patients even with endotracheal anesthesia. Therapeutic endoscopy in children is usually performed using a single-channel endoscope; however, it is technically more difficult (19).

Sclerotherapy needles for ulcer injection generally range in size from 21 to 25 G and vary in needle length from 4 to 8 mm. The 23- and 25-G needles generally are used with a 2-mm endoscopic channel; the 21-G needles are used with a 2.8-mm channel. A new double-lumen injection catheter is available that allows both hemostatic injection and simultaneous irrigation of a bleeding lesion (20). This device requires a 2.8-mm endoscopic channel. Heater probes are available in two sizes; the small, 2.4-mm probe is used with a 2.8-mm channel, and the large, 3.2-mm probe is used with a 3.7-mm channel. The bipolar electrode (BICAP) probes are also available in 2.3- and 3.2-mm sizes. Use of a laser requires at least a 2.8-mm channel.

Standard pediatric gastroduodenoscopes have a 2.0-mm channel and a 5.2- to 9.0-mm outer diameter (Table 1). These endoscopes will therefore accommodate needles for injection sclerotherapy but will not allow the use of a heater probe or BICAP. Standard adult gastroduodenoscopes have a 2.8-mm channel and an outer diameter in the range of 9.5-9.8 mm. Although their channel size is sufficient for both the small BICAP and the small heater probes, the outer diameter of these endoscopes may be prohibitive in smaller pediatric patients, especially those <2 years of age. Adult therapeutic gastroduodenoscopes have either one or two therapeutic channels, ranging in size from 2.8 to 3.7 mm, and an outer diameter in the range of 11-12.6 mm. Their larger diameter usually precludes their use in younger pediatric patients. The use of general anesthesia with a protected airway may decrease the risk of tracheal compression and subsequent hypoxia with larger endoscopes.

Pediatric colonoscopes have a 2.8-mm channel and an outer diameter ranging from 11.2 to 11.5 mm (Table 2). Adult standard and therapeutic colonoscopes have channel diameters ranging in size from 3.2 to 4.2 mm, with outer diameters ranging in size from 12.2 to 14.2 mm. The large diameter of the channel allows injection therapy, thermocoagulation, and laser therapy to be performed with pediatric colonoscopes.


Injection therapy is usually performed by injecting a hemostatic or sclerosing agent at three of four sites around an exposed bleeding vessel. The injections should be made 1-3 mm away from the vessel (Fig. 1). Maximal volumes of sclerosant have been established in adults to minimize the risk of ulcer extension or perforation. Maximum volumes of sclerosant in pediatric patients have not been studied; however, maximal adult volumes should not be exceeded. Complications including perforation may occur with even small volumes of sclerosant. Table 3 lists the most commonly used solutions and their concentrations, appropriate volumes, and estimated maximal volumes (19). Except in unusual circumstances, injection therapy should be confined to a single solution (single agent or a combination agent) during a given injection episode. Utilizing two sequential solutions increases the risk of complication with smaller volumes of sclerosant (21-23). The injection site (intravascular, perivascular, or submucosal) is specific for each agent. Without appropriate clinical trials, changing the site of injection may be hazardous. Injection therapy is not typically performed in the colon and safe volumes of sclerosant have not been determined (4). The risks of injection therapy include increased bleeding, rebleeding, bowel ischemia, and perforation (11).

The precise mechanism of the various sclerotherapeutic agents is controversial. A bleeding model using gastric serosal blood vessels in dogs suggests that a combination of factors may contribute to the formation of a hemostatic plug (24). Whittle et al. tested hemostatic solutions of normal saline, 3% hypertonic saline, epinephrine mixed with either normal or hypertonic saline, and “thrombin cocktail” [thrombin, cephapirin, and tetradecyl mixed either immediately prior to injection (fresh) or up to 3 h prior to injection (old)]. A significant decrease in the blood flow rate of transected vessels was achieved with all the solutions except for old thrombin cocktail compared to controls. The degree of histologic damage differed among the various solutions, ranging from mild hemorrhage and subserosal edema in the normal saline group to marked edema of the gastric wall with hemorrhage into the muscularis and hyaline necrosis of the muscle and blood vessels in the thrombin group.

Immediately after an injection, hemostasis is facilitated by compression and tamponade of the bleeding vessel by submucosal expansion (24,25). Epinephrine containing solutions exert an additional vasoconstrictive effect to further reduce the blood flow rate. However, this effect is transient, and therefore epinephrine may be combined with a longer acting hemostatic agent. Solutions that are hypertonic may produce tissue edema and degeneration of the vascular lumen, thereby prolonging the effects of other injected agents (11). Although not often used, the thrombin cocktail combines the thrombin effect of increased conversion of fibrinogen to fibrin, with the sclerosant effect of alcohol.

Sclerosant solutions or solutions containing ethanol may produce significant tissue damage and ulcer extension. This is particularly evident if followup endoscopy is performed within 24-48 h after injection for esophageal sclerotherapy (12,26-29). Ulceration after therapeutic injection does not appear to prolong peptic ulcer healing rates (6). There is a clear linear relation between the volume of alcohol injected and the extent of damage (30). Unlike epinephrine, the lesions produced after alcohol or sclerosant injection may be more prenounced in the submucosa than in the mucosa.

The importance of precise volumes for each agent cannot be overemphasized. The volume of absolute alcohol required to achieve hemostasis is one-tenth or less of that required with epinephrine. Injection of increased volumes of an agent may result in gastric ischemia and subsequent perforation (31,32). Injection of volumes normally considered safe within the stomach may cause perforation and or peritonitis in the thinner walled duodenum or colon (30). Injection of a combination of agents such as a sclerosant with epinephrine may decrease the maximum safe volume that may be normally used with either agent alone because of potentiation of the ischemic effects of the sclerosant by the vasoconstrictive actions of the epinephrine (21). When epinephrine is combined with a sclerosant agent to be injected sequentially, epinephrine is usually injected first in a concentration of 1:10,000 in aliquots of 1 ml per injection site (33). The volume of sclerosant injected should be based on volumes proven efficacious in clinical trials of combination regimens; when combined with epinephrine, for example, the total volume of polidocanol injection should not exceed 5-10 ml (34,35).

The ability to therapeutically inject a bleeding ulcer and achieve hemostatic control may also be limited by the location of the ulcer (36). In children undergoing therapeutic injection, the solution we prefer is epinephrine diluted with normal saline because of the decreased incidence of side effects. In some cases, tangential application of a heater probe or bipolar probe may be easier, especially along the lesser curvature of the stomach or the superior wall of the duodenal bulb. Successful injection of bleeding ulcers within other lesions, such as gastric leiomyomas, has also been reported (37).

Systemic side effects of injection therapy are a potential complication of this modality. Epinephrine absorption does appear to occur after submucosal injection, with blood levels up to five times greater than preinjection levels. This effect may be especially pronounced in patients with cirrhosis or impaired hepatic metabolism (38). Ventricular tachycardia and severe systemic hypertension have been reported after esophageal epinephrine injection (39). These side effects have not been reported with gastric or duodenal submucosal injection.


Another method of endoscopic hemostasis is via thermocoagulation, utilizing either the heater probe, monopolar, or bipolar (multipolar) coagulator. A summary of thermocoagulation techniques is listed in Table 4(19). The heater probe is composed of Teflon-coated hollow aluminum cylinders with an inner heater coil with a maximum internal temperature. The probe is water perfused to prevent tissue adherence, an advantage over monopolar coagulation, and heat is delivered via conduction to the tissue. There are small (2.4-mm)- and large (3.2-mm)-diameter probes (4). The probe is passed through the therapeutic channel. The patient should be positioned so that the blood flows away from the ulcer base if possible. In the left lateral decubitus position, blood will pool along the greater curvature of the stomach and the gastric fundus. By rolling the patient to the right side, blood will shift to the gastric cardia, allowing examination of ulcers along the greater curvature or fundus. Care must be taken to avoid aspiration of gastric contents with this maneuver (40). Heater probe coagulation is performed by initially tamponading the bleeding vessel by direct firm pressure using the heater probe and then by coagulating the vessel (Fig. 2). This technique is referred to as coaptive coagulation and is also used with the bipolar probes. The ability to compress a vessel with the probe prior to heat delivery is one reason for this technique's superior hemostatic efficacy, especially in ulcers involving large-sized arteries (41). If a twin-channel instrument is utilized, the endoscopist is able to tamponade the bleeding with the probe while simultaneously suctioning fluid from the region of the ulcer. Coagulation is usually performed in adults by two to four 30-J pulses in succession (5). In adults, the greatest success appears to be with firm tamponade on the ulcer bleeding point or nonbleeding visible vessel, then application of four pulses (total of 120 Joules) in succession before the probe position is changed (5). This technique of firm tamponade and high coagulation settings, however, increases the risk of complications when applied to other types of lesions such as Mallory Weiss tears or angiomata that traditionally occur in areas with a thinner gut wall (5). The heater probe may also be used in cases of colonic bleeding. The number of joules per pulse should be reduced, especially in right-sided colonic lesions (8,42). Although the degree of reduction has not been well studied, pulses of 15 J appear to be safe in adults (43).

The technique of utilizing a large probe with a high coagulation setting and firm pressure appears to result in a lower rebleeding rate, a decreased transfusion requirement, and a lower emergency surgery rate in randomized control studies compared to medical therapy alone and is equivalent to results achieved using multipolar electrocoagulation (5,42). A higher permanent hemostasis rate using the heater probe compared to injection therapy has been reported (44). Improved hemostasis was associated with the ease of applying the heater probe to a spurting vessel in difficult-to-reach locations. Dislodgment of the heater probe tip during therapy has been reported. In that instance, the tip was retrieved by standard biopsy forceps (45).


There are two main types of electrocoagulation probes, monopolar and bipolar. In monopolar coagulation a continuous or intermittent current is passed via the tip or side of the probe. The current is conducted to the patient's ground plate. The current is converted to high-temperature heat at the tissue contact point, which coagulates the tissue, causing collagen contraction and vessel shrinkage. For vessels <1 mm in diameter, the electrode is placed directly on the vessel and pressure is applied directly on the vessel to coapt it. With larger vessels, the coagulating current is placed circumferentially around the vessel until bleeding stops (18). Usually a midrange setting is used for 1 to 2 s per pulse at a distance of 2-3 mm from the vessel. Circumferential hemostasis is necessary because an artery in an ulcer base may bleed from either side if it is not an end artery, and therefore, a ring of tissue must be treated around the bleeding point to insure adequate hemostasis. The aim is to achieve hemostasis of the underlying artery, and not the overlying clot (4). Complications of monopolar coagulation are associated with the difficulty in regulating the depth of the burn. Deep necrosis, perforation, and delayed massive bleeding have been reported following monopolar electrocoagulation in the colon (8). Complications may be associated with electrode adherence to the underlying tissue at the treated site resulting in poor visibility (4). The tip of the probe needs to be cleaned, as coagulum accumulates during electrocoagulation (18).

Bipolar/Multipolar Coagulation

Currently, bipolar or multipolar coagulation is the method most popular with endoscopists. Bipolar coagulation is the method we most often employ to treat bleeding lesions in children. The equipment is portable and a grounding plate is not required as it is using a monopolar probe. Energy is delivered when any pair of electrodes is in contact with the bleeding target (46). The BICAP (Circon-ACMI, Stamford, CT, U.S.A.) (multipolar) probe has six points through which current can be passed; contact between any two is sufficient. This has the advantage of allowing for tangential contact, as opposed to perpendicular contact, which may be difficult because of the lesion's location. The maximal temperature achieved with this method is significantly less than that of monopolar coagulation or the Nd:Yag laser, causing less tissue injury while having greater efficacy for vessels <2 mm in diameter (41).

There are two sizes of probes available, 2.3 and 3.2 mm. Like the heater probe, the correct technique is first to compress the bleeding vessel, then to coagulate. Forceful application of the larger-size probe appears to increase the hemostatic bond strength and area and depth of coagulation (41). The greatest depth of coagulation is usually achieved with a low to midrange setting (15-25 W). Higherwatt settings produce more rapid tissue desiccation, which reduces hemostatic efficacy (47). Pulses should be applied as multiple short, 2-s pulses, or a single long pulse up to 14 s in duration (41). In adults, up to 40 s total of electrocoagulation may be required (13,41,48). The gold probe (Microvasive, Watertown, MA, U.S.A.) appears to be equivalent in efficacy to the BICAP probe but may be stiffer, allowing for greater maximal force of application (47).

Increased bleeding after bipolar coagulation has been reported in cases with a visible vessel; usually this bleeding is controllable with further bipolar coagulation, but on occasion surgery has been required. Bipolar electrocoagulation appears to be equally effective to heater probe in terms of hemostasis, incidence of rebleeding, transfusion requirement, and need for emergency surgery. Several studies report a hemostasis rate in the range of 90% for both modalities (7,14,41,49). It also appears to be as efficacious as the laser at a marked reduction in cost. The angulation of the probe or the use of a bipolar electrode compared to a laser does not appear to affect the rebleeding rate. However, poor angulation along the lesser curvature or in a deformed duodenum may make pressure application more difficult (14,49).

Angiodysplasia, predominantly of the colon, has been treated successfully by endoscopic electrocoagulation. Because of the thinner colonic wall, maximal depth of coagulation may not be desirable and therefore the probe should be applied at decreased force and for shorter intervals (47). Postpolypectomy colonic bleeding has been successfully treated with probe power settings of 3 to 5, with multiple pulses 3 to 7 s in duration (43). Hemorrhagic proctocolitis with recurrent bleeding after radiation therapy has also been successfully treated at a power setting of 5 with short, 2-s pulses by coagulating at multiple bleeding sites (50). Late rebleeding has been reported and treated successfully by repeat electrocoagulation. Occasionally, surgery may be required (51).

Laser Photocoagulation

Laser photocoagulation is another modality occasionally used to achieve endoscopic hemostasis. There are two main types of laser, the argon and the neodymium:yttrium-aluminum (Nd:Yag) laser. The argon laser's usefulness is limited because of light absorption by surrounding red blood. To use the argon laser, therefore, overlying blood must be eliminated with a coaxial air jet. Clinically, the argon laser is used primarily for right-sided colonic lesions. When compared to the Nd:Yag laser, it has a lower power and depth of tissue penetration. However, mucosal abnormalities including arteriovenous malformations absorb light energy well in the argon wavelength. The argon laser has also been used to photocoagulate polyps after biopsy in Gardner's syndrome (52).

The Nd:Yag laser is the predominant laser used in endoscopy. The laser emits a continuous wave of infrared light of a wavelength of 1,064 nm with a power up to 100 W. This light is transmitted via a 600-μm glass fiber in a 2.5-mm Teflon catheter passed down the endoscope biopsy channel. Carbon dioxide is passed coaxially along the catheter to disperse blood away from the bleeding site and to keep the fiber tip cool and free from debris. A filter is attached to the eyepiece to prevent reflected laser light from entering the endoscopist's eye. The intense laser light is directed to coagulate tissue circumferentially around the bleeding site. The recommendation using a noncontact laser in adult patients is to deliver 0.5-s pulses, at 80 W of energy, from a distance of 1 cm and at least 2 to 3 mm away from visible arterial segments for upper gastrointestinal lesions.

Application of the laser in the colon requires modification of both technique and power settings. The thermal effects of a laser beam on tissue vary according to the power density (the amount of energy converted to heat at the point where the laser beam strikes tissue) and the size of the contact area. Although the power setting and the exposure time can be preset, movement, especially in the right colon, and varying wall thickness, especially in the thin ascending colon, can change the time of exposure required to produce perforation (52). Instead of coagulating tissue, vaporization of tissue can occur. Colonic perforation secondary to laser photocoagulation is a serious risk, is more frequent in the cecum and right colon, and is more frequent with the higher-powered Nd:Yag laser. Laser burns may present with nausea, vomiting, and air in the colonic wall or free intraperitoneal air. Several authors recommend using only the argon laser at 2-8 W, for 0.5 s per application, for right-sided lesions and the Nd:Yag laser at 40-80 W, for 0.5-2.0 s per application, at a distance of 1.5-2.0 cm for left-sided lesions (52). Contact lasers allow for more efficient delivery of energy and, therefore, for reduced power settings compared to noncontact laser techniques but have not been shown to offer advantages over noncontact methods for ulcer hemostasis (53).

Lasers in colonoscopy and small bowel enteroscopy have been used for congenital vascular lesions (hereditary hemorrhagic telangiectasia and blue rubber bleb nevi syndrome) as well as superficial vascular lesions including angiodysplasia, telangiectasias, and arteriovenous malformation (7,54,55). Asymptomatic nonbleeding angiodysplasias are not treated (8). One problem after laser photocoagulation of a lesion is that the histologic diagnosis is difficult to confirm (4,52). Like the heater probe, laser therapy can also provoke bleeding, which can usually be stopped with further laser coagulation (55). In addition, there is an increased chance of full-thickness perforation of the gut wall (56). There has been a case report of the laser being utilized to treat congenital fine vascular lesions in the stomach of an 11-year-old boy, and with further study the laser may be utilized primarily for such lesions in the future (56).

There is a long learning curve associated with use of the laser and this modality should be utilized only by experienced operators. Currently, the laser seems to offer little advantage over the heater probe and BICAP, and because of its increased cost and decreased portability, the other two modalities are likely to predominate in the foreseeable future.

Hot Biopsy Forceps

An additional coagulation technique used primarily in the colon is the hot biopsy forceps. The size of the forceps requires a 2.8-mm endoscopic channel (57). The forceps are used to tent the mucosa lifted away from the colonic muscular layer. A brief electrocoagulating current passes through the forceps to the mucosa, causing coagulation at its base while preserving the histologic integrity of the specimen. The unit is set on coagulation, with no cutting, at a setting of 10-15 W for 1-2 s in the cecum and ascending colon and up to 15-20 W for 2 s in the left colon. Small angiomata of the colon may be coagulated using a similar technique for 1-2 s (58). Higher settings or longer application times have been associated with an increased risk of perforation, especially in the right colon (59,60).

Perforation has also been reported with application of the technique to the upper gastrointestinal tract (57-59). In the colon the risk of perforation is 0.05%; this is less than the complication rate associated with snare polypectomy but greater than that of routine colonoscopy (60,61).


The choice of therapeutic endoscopic technique depends on the training and equipment available to the endoscopist. If the technique is properly performed, the results are similar using injection, thermal coagulation, or laser therapy. We recommend that pediatric endoscopists concentrate on one thermal and one injection technique, since individual bleeding lesions may be more amenable to one method than another based on their anatomic location or briskness of bleeding.

FIG. 1.
FIG. 1.:
Endoscopic injection of a sclerosing solution around a visible vessel. Because the vessel is not an end artery, therapeutic injection is performed circumferentially around the vessel.
FIG. 2.
FIG. 2.:
Coaptive coagulation using a heater probe. A and B: Initially the exposed vessel is tamponaded by direct firm tangential pressure utilizing the heater probe until blanching of the vessel occurs. C: The heater probe is then used to coagulate the blood vessel. D: After coagulation, the exposed vessel appears blanched.


1. Ament ME, Christie DL. Upper gastrointestinal fiberoptic endoscopy in the pediatric patient. Gastroenterology 1977;72:1244-8.
2. NIH Consensus Development Conference. Therapeutic endoscopy and bleeding ulcers. JAMA 1989;262:1369-72.
3. Kovacs TO, Jensen DM. Endoscopic control of gastroduodenal hemorrhage. Annu Rev Med 1987;38:267-77.
4. Sugawa C. Endoscopic diagnosis and treatment of upper gastrointestinal bleeding. Surg Clin North Am 1989;69(6):1167-83.
5. Jensen DM. Heat probe for hemostasis of bleeding peptic ulcers: techniques and results of randomized controlled trials. Gastrointest Endosc 1990;36(5):S42-9.
6. Standards of Practice Committee, American Society for Gastrointestinal Endoscopy. The role of endoscopy in the management of non-variceal acute upper gastrointestinal bleeding. Gastrointest Endosc 1992;38:760-4.
7. Jensen DM, Machicado GA, Kovacs TO, et al. Controlled randomized study of heater probe and BICAP for hemostasis of severe ulcer bleeding. Gastroenterology 1988;97(5):A208.
8. Schrock TR. Colonoscopic diagnosis and treatment of lower gastrointestinal bleeding. Surg Clin North Am 1989;69(6):1309-25.
9. Gilbert DA, Silverstein FE. Endoscopy in gastrointestinal bleeding. In: Sivak MV Jr, ed. Gastroenterologic endoscopy. Philadelphia: Saunders, 1987:110-27.
10. Monahan DW, Peluso FE, Goldner F. Combustible colonic gas levels during flexible sigmoidoscopy and colonoscopy. Gastrointest Endosc 1992;38:40-3.
11. Hirao M, Kobayashi T, Masuda K, et al. Endoscopic local injection of hypertonic saline-epinephrine solution to arrest hemorrhage from the upper gastrointestinal tract. Gastrointest Endosc 1985;31(5):313-7.
12. Lazo MD, Andrade R, Medina MC, et al. Effect of injection sclerosis with alcohol on the rebleeding rate of gastroduodenal peptic ulcers with nonbleeding visible vessels: a prospective controlled trial. Am J Gastroenterol 1992;87:843-6.
13. Laine L. Multipolar electrocoagulation in the treatment of peptic ulcers with nonbleeding visible vessels. Ann Intern Med 1989;110:510-4.
14. Jensen D, Hirabayshi K. A study of coagulation depths with bicap and heater probe to improve endoscopic hemostasis of bleeding peptic ulcers. Gastrointest Endosc 1989;25(2):181.
15. Matthewson K, Swain CP, Bland M, et al. Randomized comparison of Nd: Yag laser, heater probe, and no endoscopic therapy for bleeding peptic ulcers. Gastroenterology 1990;98:1239-44.
16. Waring JP, Sanowski RA, Sawyer RL, et al. A randomized comparison of multipolar electrocoagulation and injection sclerosis for the treatment of bleeding peptic ulcer. Gastrointest Endosc 1991;37(3):295-8
17. Chung SC, Leung JW, Sung JY, et al. Injection or heat probe for bleeding ulcer. Gastroenterology 1991;100(1):33-7.
18. Papp JP. Monopolar and electrohydrothermal treatment of upper gastrointestinal bleeding. Gastrointest Endosc 1990;36(5):S34-7.
19. Wyllie R, Kay M. Esophagogastroduodenoscopy, colonoscopy and related techniques. In: Wyllie R, Hyams J, eds. Pediatric gastrointestinal disease: pathophysiology, diagnosis, management. Philadelphia: Saunders, 1993:967-98.
20. Laine L. A new double lumen, direct irrigation injection therapy catheter. Gastrointest Endosc 1992;38:594-6.
21. Loperfido S, Patelli G, La Torre L. Extensive necrosis of gastric mucosa following injection therapy of bleeding peptic ulcer. Endoscopy 1990;22:285-6.
22. Chung SC, Leung JW, Leong HT, et al. Adding a sclerosant to endoscopic injection in actively bleeding ulcers: a randomized trial. Gastrointest Endosc 1993;39:611-5.
23. Saeed ZA, Winchester CB, Michaletz PA, et al. A scoring system to predict rebleeding after endoscopic therapy of nonvariceal upper gastrointestinal hemorrhage, with comparison of heat probe and ethanol injection. Am J Gastroenterol 1993;88:1842-9.
24. Whittle TJ, Sugawa C, Lucas CE, et al. Effect of hemostatic agents in canine gastric serosal vessels. Gastrointest Endosc 1991;37(3):305-9.
25. Fuchs KH, Wirtz HJ, Elfeldt R. Initial experience with thrombin as injection agent for bleeding gastroduodenal lesions. Endoscopy 1986;18:146-8.
26. Chen PC, Wu CS, Liaw YF. Hemostatic effect of endoscopic local injection with hypertonic saline-epinephrine solution and pure ethanol for digestive tract bleeding. Gastrointest Endosc 1986;32:319-23.
27. Hassall E, Berquist WE, Ament ME, Vargas J, Dorney S. Sclerotherapy for extrahepatic portal hypertension in childhood. J Pediatr 1989;115:69-74.
28. Hill ID, Bowie MD. Endoscopic sclerotherapy for control of bleeding varices in children. Am J Gastroenterol 1991;86:472-6.
29. Howard ER, Stringer MD, Mowat AP. Assessment of injection sclerotherapy in the management of 152 children with oesophageal varices. Br J Surg 1988;75:404-8.
30. Rutgeerts P, Geboes K, Vantrappen G. Experimental studies of injection therapy for severe nonvariceal bleeding in dogs. Gastroenterology 1989;97:610-21.
31. Bedford RA, van Stolk R, Sivak MV Jr, Chung RS, Van Dam J. Gastric perforation after endoscopic treatment of a Dieulafoy's lesion. Am J Gastroenterol 1992;87:244-7.
32. Chester JF, Hurley PR. Gastric necrosis: a complication of endoscopic sclerosis for bleeding peptic ulcer. Endoscopy 1990;22:287.
33. Ruzkowski CJ, Sanowski RA. Injection therapy for hemostasis of bleeding peptic ulcers. Gastrointest Endosc Clin North Am 1991;1:303-18.
34. Soehendra N, Grimm H, Stenzel M. Injection of nonvariceal bleeding lesions of the upper gastrointestinal tract. Endoscopy 1985;17:129-32.
35. Panes J, Forne M, Bagena F, Viver J. Endoscopic sclerosis in the treatment of bleeding peptic ulcers with a visible vessel. Am J Gastroenterol 1990;85:252-4.
36. Lin HJ, Pering CL, Lee FY, et al. Endoscopic injection for arrest of peptic ulcer hemorrhage: final results of a prospective randomized comparative trial. Gastrointest Endosc 1993;39:15-9.
37. Giorcelli W, Rodi M. Injection sclerotherapy for bleeding gastric leiomyoma. Gastrointest Endosc 1992;38:730-1.
38. Sung JY, Chung SC, Low JM, et al. Systemic absorption of epinephrine after endoscopic submucosal injection in patients with bleeding peptic ulcers. Gastrointest Endosc 1993;39:20-2.
39. Stevens PD, Lebwhol O. Hypertensive emergency and ventricular tachycardia after endoscopic epinephrine injection as a Mallory-Weiss tear. Gastrointest Endosc 1994;40:77-8.
40. Freeman ML. The current endoscopic diagnosis and intensive care unit management of severe ulcer and other nonvariceal upper gastrointestinal hemorrhage. Gastrointest Endosc Clin North Am 1991;1:209-39.
41. Laine L. Bipolar/multipolar electrocoagulation. Gastrointest Endosc 1990;36(5):S38-41.
42. Jenson DM, Machicado GA. Diagnosis and treatment of severe hematochezia. Gastroenterology 1988;95:1569-74.
43. Rex DK, Lewis BS, Waye JD. Colonoscopy and endoscopic therapy for delayed post polypectomy hemorrhage. Gastrointest Endosc 1992;38:127-9.
44. Lin HJ, Tsai YT, Lee SD, et al. A prospective randomized trial of heat probe thermocoagulation versus pure alcohol injection in nonvariceal peptic ulcer hemorrhage. Am J Gastroenterol 1988;83(3):283-6.
45. Meichner RH, Galambos J. Heater probe dislodgement during use. Am J Gastroenterol 1993;88:151-2.
46. Yamamoto H, Hajiro K, Matsui H, et al. Endoscopic bipolar electrocoagulation: development of a new bipolar coagulator for stopping gastrointestinal bleeding. Gastroenterol Jpn 1982;17(2):75-9.
47. Laine L. Determination of optimal technique for bipolar electrocoagulation treatment. Gastroenterology 1991;100:107-12.
48. Laine L. Multipolar electrocoagulation in the treatment of active upper gastrointestinal tract hemorrhage. N Engl J Med 1987;316(26):1613-7.
49. Hui WM, Ng MM, Lok AS, et al. A randomized comparative study of laser photocoagulation, heater probe, and bipolar electrocoagulation in the treatment of actively bleeding ulcers. Gastrointest Endosc 1991;37(3):299-304.
50. Mauhoury V, Brunetaud JM, Cortot A. Bipolar electrocoagulation treatment for hemorrhagic radiation injury of the lower digestive tract. Gastrointest Endosc 1991;37(4):492-3.
51. Rogers BH. Endoscopic electrocoagulation of vascular abnormalities of the gastrointestinal tract in 51 patients. Gastrointest Endosc 1982;28(2):142-3.
52. Buchi KN. Endoscopic laser surgery in the colon and rectum. Dis Colon Rectum 1988;31:739-45.
53. Swain CP. Nd:YAG laser for treatment of bleeding ulcers. Gastrointest Endosc Clin North Am 1991;1:341-66.
54. Shahed M, Hagenmuller F, Rosch T, et al. A 19-year-old female with blue rubber bleb nevus syndrome. Endoscopy 1990;22:54-6.
55. Lanthier PH, D'Harveng B, Vanheuverzwyn R, et al. Colonic angiodysplasia. Dis Colon Rectum 1989;32:296-8.
56. Noronha P, Leist M. Endoscopic laser therapy for gastrointestinal bleeding from congenital vascular lesions. J Pediatr Gastroenterol Nutr 1988;7(3):375-8.
57. Williams, CB. Small polyps: the virtues and the dangers of hot biopsy. Gastrointest Endosc 1991;37(3):394-5.
58. Gilbert DA, DiMarino AJ, Jensen DM, et al. Status evaluation: hot biopsy forceps. Gastrointest Endosc 1992;38:753-6.
59. Wadas DD, Sanowski RA. Complications of the hot biopsy forceps technique. Gastrointest Endosc 1988;34(1):32-7.
60. Quigley EM, Donovan JP, Linder J, et al. Delayed massive hemorrhage following electrocoagulating biopsy (“hot biopsy”) of a diminutive colonic polyp. Gastrointest Endosc 1989;35(6):559-63.
61. Nelson AM. Delayed hemorrhage following “hot biopsy” of a diminutive colonic polyp. Gastrointest Endosc 1990;36:418.
© Lippincott-Raven Publishers