George Gershman, Jorge H. Vargas, and Mike Thomson Acute gastrointestinal bleeding (AGIB) in children is a true medical emergency associated with significant morbidity and mortality even in developed countries. Recent retrospective analysis of AGIB in patients admitted to the 47 pediatric tertiary centers in the US (data extracted from Pediatric Hospital Information System) revealed mortality of almost 0.4% and 3% in children with the primary and secondary diagnosis of AGIB respectively. The key preventable reasons why children are still dying from AGIB are underestimation of severity of bleeding on admission, delayed endoscopy, lack of specialists trained in therapeutic endoscopy, and miscalculation of risk of recurrent bleeding. To address the first pair of issues, the Sheffield scoring system has been developed to select children who will benefit from urgent endoscopy and endoscopic hemostasis. The score is computed by adding the numerical value for clinically relevant characteristics in four categories: A cut‐off of 8 for intervention from a maximal score of 24 has a positive predictive value and negative predictive value of about 91% and 89% respectively. The Sheffield scoring system stratifies children with AGIB in three categories: 1 immediate intervention for children with uncontrolled bleeding who require volume control; 2 endoscopic intervention within 12 hours for children in whom the threshold score is reached, but who are stable; 3 elective or no endoscopy for children whose clinical bleeding risk score does not reach the intervention threshold and in whom AGIB would appear to have ceased. The next challenge is a lack of equipment and specialists in endoscopic hemostasis in children. One answer to this could be development of AGIB “super‐centers” or collaboration with adult gastroenterologists skilled in endoscopic hemostasis. Each approach has pros and cons based on differences in hospital policies, type of medical systems (social versus private), and lack of enthusiasm on the part of our adult colleagues to get involved in the pediatric field. It has been proved that early endoscopy within 24 hours is associated with reduction in transfusions, rebleeding (RB) and need for surgery compared to later endoscopy. Furthermore, endoscopic stigmata of active or resent bleeding organized by Forrest classification (F) defines indications for endoscopic hemostasis and hospitalization based on the probability of rebleeding risk: In addition to the Forrest classification, there are other endoscopic features of peptic ulcers that can predict adverse outcomes and/or endoscopic treatment failure. These include large ulcer (>2 cm), large nonbleeding visible vessel, presence of blood in the gastric lumen, and ulcer location on the posterior duodenal wall or proximal lesser curvature of the stomach. Further research is needed to prove the same concept in children. Low incidence of non‐variceal AGIB in children and lack of pediatric research are the reasons why pediatric gastroenterologists rely on recommendations derived from the practice of our adult colleagues. The main source for diagnosis and management of non‐variceal AGIB is the guidelines from the American College of Gastroenterology and European Society of Gastrointestinal Endoscopy. The key elements of these recommendations acceptable to pediatric practice are: Endoscopy during active bleeding is challenging due to poor visibility and the need to remove large amounts of blood and manipulate with a hemostatic device and flushing water on the target simultaneously. Therefore, a double‐channel therapeutic endoscope is preferable if it is appropriate to the patient size. If this is not feasible, an adult‐size upper GI endoscope can be used in children more than 10 kg. Pediatric‐size upper GI endoscopes have limited suction capacity and are reserved for toddlers. Slim endoscopes ≤6 mm are kept for neonates and infants. Methods of endoscopic hemostasis can be classified into three categories: Epinephrine in saline (1:10 000) is a vasoconstrictive agent routinely used for hemostasis in children with non‐variceal bleeding. It is injected into the bleeding site through the 23 or 25 gauge needle (both are available for endoscopes with 2.8 mm and 2 mm biopsy channels). The needle should be primed (filled with epinephrine) before insertion into the biopsy channel to prevent air embolism. Currently, injection of epinephrine is recommended in combination with thermal or constrictive endoscopic devices, particularly in patients with bleeding ulcer (Forrest 1a, 1b, and 2a). Limited pediatric data reveal a high rate of rebleeding after epinephrine monotherapy (40%) in children with acute non‐variceal upper gastrointestinal bleeding, supporting a concept of combination endoscopic hemostasis. Indications for epinephrine injections are: The injection technique should be adjusted to the specific cause of bleeding; for example, a bleeding ulcer with a visible vessel requires peripheral four quadrants injections, typically 0.5–1 mL per site, followed by direct injection into the bleeding vessel. In contrast, target injection into the bleeding spot is an initial step of endoscopic hemostasis caused by ateriovenous malformations or bleeding from the base of the polyp after polypectomy. The amount of epinephrine solution should not exceed 16 mL (recommended adult dose). In our practice, we rarely inject more than 4 mL of epinephrine solution per session to avoid local ischemia or perforation. Topical application of hemostatic agents is a new, nontraumatic, and very effective type of endoscopic hemostasis which has been rapidly gaining ground in adult gastroenterology over the last decade. It can be used for localized or diffuse lesions as primary or salvage treatment, when bleeding persists despite application of conventional methods of endoscopic hemostasis. Three agents are commercially available in the United States. Metal clips are very effective in controlling active hemorrhage from bleeding vessels in a peptic ulcer base, Dieulafoy lesion, Mallory–Weiss tears, and polypectomy site. Single‐use preloaded rotatable two‐pronged clips, which can be reopened and repositioned up to five times, are commercially available (Figure 30.5). The endoscope with 2.8 mm biopsy channels is necessary to accommodate a standard clipping device, which consists of a metal cable within a metal coil sheath covered by a 2.2 mm Teflon® catheter. However, application of a Resolution Clip® (Boston Scientific) without a Teflon catheter allows it to adapt to an endoscope with a smaller 2.2 mm working channel. The keys for successful application of a clipping device are: The most challenging scenario for hemostatic clip therapy is bleeding from a large ulcer on the posterior wall of the duodenal bulb or superior duodenal angle. Rotating the patient into prone or supine and use of a double‐channel therapeutic endoscope may improve visibility and navigation of the scope in front of the bleeding vessel to secure the optimal conditions for clip deployment. Primary hemostasis of a bleeding ulcer or Dieulafoy lesion can be achieved in 84–100% of cases respectively. Complications associated with clipping hemostasis are quite rare. Since 2007, over‐the‐scope clips (OTSC) have been successfully used by adult gastroenterologists for endoscopic hemostasis and closure of perforations (Figures 30.6 and 30.7). In conjunction with its sister device, it allows for full‐thickness biopsies of the low GI tract. It is available in three different sizes: 11, 12, and 14 mm with 3 or 6 mm depths of caps for grasping more or less tissue during approximation. The OTSC is mounted over the tip of the endoscope with an applicator similar to a cup for an esophageal varices ligature. It may be used in combination with an additional tool such as a “twin grasper” to approximate edges before deployment of the clip. The device is controlled by an extraendoscopic channel which is attached to the outside of the scope. The central grasping forceps are deployed to stabilize the tissue of choice and then suction is applied for five seconds until vision is obliterated, similar to variceal banding, and then the jaws of the capture clip are released to oppose each other for closure when the handle is turned outside the patient. Thermal hemostasis embraces different methods which target non‐variceal causes of bleeding, such as: Three types of thermal devices are currently used in pediatric practice. The heating unit of these devices consists of two (bipolar – bipolar probe is also called the ‘gold’ probe and can have a needle within it in order to inject epinephrine quadrantically around a point bleeder in order to achieve hemostasis before application of the gold probe to the visible vessel) or 4–6 (multipolar) active electrodes incorporated into a thermal probe. Advantages of the system are: Small (2.2 mm) and large (3.2 mm) bipolar or multipolar probes are commercially available. Large probes allow for application of stronger pressure on a bleeding vessel and depth of coagulation. The depth of coagulation is related to the power setting. Low‐to‐mid range of setting (15–25 W) is preferable for deep coagulation. Escalating the power setting increases water evaporation leading to a diminished degree of coagulation. The device generates and controls heat up to 250 °C by pulses of energy delivered to a silicon clip surrounded by a low heat capacity metal envelope without any electric current in the tissue. The probe is supplemented by a three‐jet water system. The metal envelope warms up to the designated temperature in less than 0.2 seconds and cools off in less than 0.5 seconds. The computer controls the temperature and total energy delivered to the tissue. The endoscopist adjusts the energy delivery to the source of bleeding in the range of 5–30 J. Advantages of the heater probe include: Bipolar/multipolar and heater probes have been used more often in pediatric patients than any other type of thermal hemostatic devices. Commercially available bipolar/multipolar and heater probes fit easily into the 2.8 mm biopsy channel endoscopes. Both methods provide enough heat for coagulation of mesenteric arteries up to 2 mm in experimental models. Plasma coagulation is the result of ionization of a noble gas (argon is the cheapest), which fills a small gap between the electrical electrode and the target tissue. Ionization of argon occurs when a high‐frequency current creates sufficient electric field strength. Ionized argon conducts an electrical current and flows along the same pathway. The released energy induces desiccation and coagulation without carbonization and evaporation, which prevents deep tissue destruction. The depth of coagulation is proportional to the power setting and application time but almost never exceeds 3 mm. Holding a probe in one site for five seconds produces coagulation depth of 2–3 mm with a power setting of 30–60 W. The advantages of APC coagulation are: The procedure carries a risk of perforation due to direct contact between tissue and probe, and stretching of the stomach and bowel wall due to the accumulation of argon. Thin (1.5 mm) probes are commercially available and suitable for small‐caliber pediatric endoscopes. This makes it possible to apply APC even in neonates and infants. Two types of complications have been described in adults: perforation or submucosal emphysema due to direct contact of the probe with mucosa and flow of argon gas through the damaged mucosa. A detailed description of endoscopic hemostasis with different thermal devices is beyond the scope of the chapter. Before the procedure, the pediatric gastroenterologist should become familiar with the available equipment, the proper setting of the coagulator, and optimal treatment requirements for different types of bleeding lesions. Bleeding from a non‐variceal lesion in the stomach or duodenum can be arterial from a visible vessel or venous/parenchymal. During endoscopy, a visible nonbleeding vessel appears as a pyramid‐like protuberance in the ulcer base. Observation of a forceful pulsating eruption of blood from the ulcer leaves no doubt about the arterial nature of the bleeding. Immediate endoscopic intervention is required, creating pressure by forcing a bipolar or heater probe up against the bleeding vessel, followed by four pulses of 30 J using the heater probe or 8–10‐second pulses with a power setting of 15–20 W on a 50 W generator for bipolar or multipolar probe before repositioning the probe. Coagulation is repeated until the visible vessel becomes flat and bleeding is stopped. The procedure is complicated by poor visibility, especially when the source of bleeding is located in the duodenum. Forceful irrigation and suction is more effective with therapeutic endoscopy. Finding a large blood clot at the site of bleeding is another challenging situation. The risk of worsening bleeding has to be weighed against the potential benefits before an attempt to remove an adherent clot to expose a bleeding vessel. Treatment requires careful washing out of blood and loose fibrin until the edge of the ulcer becomes visible. The next step is injection of epinephrine under the clot in a circular fashion, to decrease the risk of bleeding during dislodgment of the clot from the ulcer and expose the underlying vessel for endoscopic hemostasis. Once again, this is a challenging procedure, which requires a highly skillful endoscopist and supportive team. Presence of a nonbleeding vessel permits better assessment of the lesion and more precise positioning of the hemostatic device. The power setting and force application should be adjusted in patients with angiodysplasia to avoid perforation. The APC technique is preferable in such circumstances.
30
Non‐variceal endoscopic hemostasis
Introduction
General considerations
Choice of endoscope
Techniques of endoscopic hemostasis
Epinephrine injection therapy
Endoscopic hemostatic powder and gel
Hemostatic clips
Over‐the‐scope clip
Thermal coagulation
Bipolar or multipolar thermal devices
Computer‐controlled thermal probes (heater probes)
Argon plasma coagulation (APC)
Technique of thermal coagulation