Endoscopy

Gastrointestinal endoscopy is the direct or instrument-assisted examination of the gastrointestinal tract and adjacent pancreatobiliary structures. It combines visualization with tissue sampling, intraluminal imaging, lesion characterization, and treatment; the same examination may therefore be diagnostic, surveillance-based, staging-oriented, or therapeutic. Its value is greatest when these functions are matched to the question being asked: visualization detects and localizes disease, biopsy supplies histopathologic confirmation, cross-sectional or ultrasound imaging defines depth and extraluminal involvement, and intervention can remove, drain, dilate, stent, or otherwise treat the lesion. Endoscopic procedures are fundamental to both diagnosis and management of gastrointestinal disorders ^[1].

Esophagogastroduodenoscopy (EGD), also called upper gastrointestinal endoscopy, examines the esophagus, stomach, and duodenum. It permits targeted biopsy and endoscopic treatment of mucosal disease, including resection or ablation of selected neoplasia. EGD may be used for initial diagnosis, for surveillance after endoscopic treatment, or for local assessment of a lesion whose depth and morphology determine whether endoscopic therapy is appropriate. Sedated EGD remains the standard screening method for Barrett’s esophagus, although its invasiveness, sampling error, cost, and procedural risks limit population-wide use ^[15].

Colonoscopy examines the colon and usually the terminal ileum. Unlike tests that only image the bowel, it permits biopsy and immediate treatment, including polypectomy and endoscopic resection; colonoscopy is therefore both a diagnostic test and a therapeutic procedure ^[3]. It also provides surveillance after polyp removal or treatment of colorectal neoplasia, although surveillance intensity should reflect absolute cancer risk and the burden imposed on patients and endoscopy services ^[17]. Flexible sigmoidoscopy reaches the rectum and distal colon. Its narrower anatomic field makes it useful for distal mucosal assessment, biopsy, and selected local therapy, but it cannot substitute for examination of the proximal colon when disease there remains clinically relevant.

Enteroscopy extends direct endoscopic examination beyond the reach of standard EGD and colonoscopy. Push enteroscopy generally reaches the proximal small bowel, whereas device-assisted enteroscopy , using balloon or spiral assistance, can examine substantially deeper small-bowel segments and allows biopsy, hemostasis, dilation, and selected lesion therapy. These procedures are distinct from capsule endoscopy , in which an ingestible camera records images while traversing the gastrointestinal tract. Capsule endoscopy is primarily a mucosal imaging and localization test: it expands small-bowel visualization without providing tissue, insufflation, irrigation, or immediate treatment. In a prospective multicenter study of iron-deficiency anemia, small-bowel capsule endoscopy identified potential bleeding lesions more often than either EGD or colonoscopy, illustrating its complementary rather than interchangeable role ^[13].

Endoscopic ultrasonography (EUS) combines endoscopic access with high-frequency ultrasound from within the gastrointestinal lumen. Its defining contribution is imaging beyond the mucosal surface: it characterizes the layers of the bowel wall, adjacent lymph nodes, and organs such as the pancreas and biliary tree, while allowing needle-based tissue acquisition and selected drainage procedures. EUS is consequently used for lesion characterization and locoregional staging when mucosal inspection alone cannot establish depth or relationship to surrounding structures. AI applications in EUS, ERCP, and cholangioscopy are promising, but much of the available literature remains single-center or retrospective, so reported performance should not be treated as equivalent to routine clinical benefit ^[4].

Endoscopic retrograde cholangiopancreatography (ERCP) is a fluoroscopy-assisted endoscopic procedure that accesses the major papilla through a duodenoscope to image and treat the bile and pancreatic ducts. In contemporary practice its principal function is therapeutic rather than purely diagnostic: stone extraction, sphincter therapy, stricture dilation, and biliary or pancreatic stent placement are performed during ductal access. Cholangioscopy , usually digital single-operator cholangiopancreatoscopy performed through an ERCP platform, adds direct visualization of the bile or pancreatic duct and permits targeted assessment and treatment of complex ductal disease; it is described as an adjunct to ERCP for diagnosis and treatment ^[7]. Endoscopic papillary large-balloon dilation combined with sphincterotomy is an example of ERCP-based therapy for difficult biliary stones ^[6].

Procedure-specific platforms broaden what can be done through an endoscope. Endoscopic mucosal resection (EMR) and endoscopic submucosal dissection (ESD) remove neoplasia confined to, or extending through, the mucosa and submucosa; ESD is designed for en bloc resection and can reduce recurrence compared with EMR for large colorectal polyps, although selection depends on morphology and invasion risk ^[16]. Endoscopic full-thickness resection extends resection through the bowel wall for selected complex colorectal lesions, including non-lifting lesions, and may offer a minimally invasive alternative to surgery; histopathology remains necessary to determine the completeness and oncologic risk of the resection ^[8]. A diagnostic endoscopic dissection strategy can also provide definitive depth assessment before committing selected patients to surgery ^[10]. Endoscopic bariatric techniques use therapeutic devices or suturing platforms to modify gastric volume or anatomy and belong to the therapeutic, rather than purely visualizing, arm of the discipline.

Endoscopy does not replace radiology , surgery , or histopathology . Radiology supplies a broader view of the bowel, solid organs, ducts, mesentery, and distant disease; endoscopy supplies surface detail, tissue, and immediate intervention. Histopathology determines cellular diagnosis, grade, depth, margin status, and other features that govern treatment, but the pathologist’s interpretation is constrained by the tissue obtained and by the endoscopist’s description of the lesion and resection. Surgery remains necessary when disease extends beyond safe or curative endoscopic boundaries, when lymph-node clearance is required, or when endoscopic therapy cannot achieve durable control. The disciplines therefore function as complementary components of a diagnostic and therapeutic pathway rather than as competing versions of the same test.

The practical classification is therefore functional as well as anatomic. A diagnostic procedure answers what and where; a surveillance procedure searches for recurrence or metachronous disease at defined intervals; a staging procedure estimates depth, nodal involvement, or ductal extension; and a therapeutic procedure changes the disease course during the examination. A single session may serve more than one purpose, but the distinction matters because tissue acquisition, cross-sectional imaging, multidisciplinary review, and definitive surgery may still be required even when the endoscope provides the initial diagnosis or treatment.

Choose the examination that can answer the clinical question and act on the answer during the same encounter. The decision rests on four variables: the suspected lesion, the anatomic territory, the urgency of possible treatment, and the patient’s pretest probability. A test that can only demonstrate disease is inadequate when the likely finding requires hemostasis, dilation, stenting, tissue acquisition, or resection. Conversely, endoscopy is a poor first test when the suspected process is predominantly extraluminal, when cross-sectional staging will determine management, or when a positive finding would not alter treatment.

Upper gastrointestinal indications

Perform EGD early when the suspected abnormality is mucosal or intraluminal in the esophagus, stomach, or duodenum. Progressive dysphagia warrants EGD because direct inspection can distinguish inflammatory, malignant, and structural disease while permitting biopsy and dilation; complete obstruction or an impacted bolus requires an urgent therapeutic examination rather than a purely diagnostic study. In suspected achalasia, endoscopy excludes an obstructing lesion and pseudoachalasia, whereas fluoroscopy and physiologic testing answer complementary questions about lumen configuration and motor function. Endoscopic myotomy is an established treatment for achalasia, but the choice of therapy depends on the confirmed diagnosis and the patient’s anatomy rather than symptoms alone. ^[35]

For upper gastrointestinal bleeding, plan the index EGD around the possibility of hemostasis. If active bleeding is likely, a diagnostic study that cannot inject, clip, coagulate, or otherwise control the lesion creates a predictable second procedure. Use CT angiography instead when bleeding is brisk or intermittent in a way that makes endoscopic visualization unlikely, when the source may be beyond the reach of EGD, or when cross-sectional localization will determine embolization or surgery. Do not use EGD to investigate nonspecific abdominal symptoms when the pretest probability of an actionable upper gastrointestinal lesion is low and noninvasive evaluation would not change management.

Lower gastrointestinal indications

Choose colonoscopy when the relevant question includes proximal colonic disease, colorectal cancer, inflammatory bowel disease, iron-deficiency anemia, unexplained rectal bleeding, or a need for biopsy or polypectomy. Flexible sigmoidoscopy is reasonable only when symptoms and prior information localize the concern to the rectum or distal colon and the probability of proximal disease is low. A normal sigmoidoscopy cannot exclude a proximal cancer, right-sided bleeding source, or ileocolonic inflammatory disease. CT colonography can assess the colonic lumen when colonoscopy is incomplete or unsuitable, but a positive result still requires colonoscopy for biopsy or removal.

Screening and surveillance examinations should be offered when detection of a premalignant lesion or early cancer can alter outcome. Colonoscopy is preferred when the patient accepts an invasive examination and immediate polypectomy has value; nonendoscopic screening is reasonable when participation is more likely with a less invasive strategy. The examination is not justified by the label “screening” alone: estimate life expectancy, competing illness, prior findings, and whether treatment of a detected lesion would be pursued. In a patient who would not undergo cancer treatment or polypectomy, screening endoscopy is unlikely to improve outcomes.

Small-bowel evaluation

After negative EGD and colonoscopy in a patient with ongoing or recurrent bleeding, capsule endoscopy is usually the least invasive way to inspect the small-bowel mucosa. Its strength is broad mucosal coverage without intubation; its limitation is the absence of biopsy, irrigation, insufflation, and immediate treatment. Capsule endoscopy therefore localizes a target rather than completing the intervention. Use device-assisted enteroscopy when the expected lesion requires biopsy, hemostasis, dilation, tattooing, or resection, and select the insertion route from the lesion’s estimated location.

Do not begin with capsule endoscopy when obstruction, a tight stricture, a large infiltrating mass, or a predominantly extraluminal process is plausible. CT enterography or MR enterography is better for mural thickening, transmural inflammation, fistula, abscess, mass effect, and stenosis; MR enterography avoids ionizing radiation when repeated cross-sectional assessment is anticipated. A negative capsule examination is less informative when the suspected disease is extraluminal or when the clinical question concerns surgical anatomy rather than mucosal lesions. Machine-learning performance on capsule images remains an adjunct to expert review: a recent benchmark used image-level rather than patient-level splitting, so its results should not be treated as patient-level clinical accuracy. ^[31]

EUS and pancreaticobiliary endoscopy

Use EUS when the central question concerns the bowel wall, a subepithelial lesion, an extrinsic impression, the pancreas, bile duct, gallbladder, or regional nodes. EUS can define whether a subepithelial lesion arises from the mucosa, submucosa, or muscularis propria and can obtain targeted tissue; CT or MRI is preferred when the principal need is assessment of vascular encasement, distant metastases, multifocal disease, or a large extraluminal mass. EUS is most valuable when its result will change staging, tissue diagnosis, resectability, or treatment selection. EUS-guided liver biopsy and portal-pressure measurement are expanding applications, but these advanced procedures should be concentrated in expert services because their value depends on technical expertise and integration with hepatobiliary management. ^[21]

Use MRCP or contrast-enhanced CT to establish the cause and extent of suspected biliary or pancreatic obstruction when immediate intervention is not required. MRCP provides noninvasive ductal mapping; CT better defines a mass, vascular anatomy, and distant disease. Go directly to ERCP when the probability of obstruction is high and the patient needs drainage, stone extraction, sphincter therapy, stricture dilation, or stent placement. ERCP should not be used as a diagnostic substitute for EUS or MRCP when no intervention is anticipated, because its principal value is therapeutic. Contemporary biliary practice also includes EUS-guided drainage after unsuccessful or unsuitable ERCP, but comparative evidence across endoscopic, percutaneous, and surgical drainage strategies remains unsettled and is the subject of ongoing comparative research. ^[22]

When endoscopic access is impossible or drainage cannot be achieved safely, choose percutaneous drainage or surgery according to anatomy, resectability, infection, prognosis, and the patient’s goals. Do not persist with repeated diagnostic procedures when cross-sectional imaging demonstrates a lesion requiring oncologic resection, lymph-node clearance, or management of perforation, ischemia, diffuse peritonitis, or uncontrolled sepsis. Endoscopic palliation is appropriate when it can relieve obstruction or bleeding with less physiologic burden than surgery and when the patient would benefit from symptom control; it is inappropriate when the expected benefit is too small to justify procedural risk or when definitive surgery is already required.

Pearl: Match the modality to the lesion and to the action that must follow it: use endoscopy for reachable mucosal or luminal disease and immediate therapy, cross-sectional imaging for mural, extraluminal, vascular, or distant disease, and surgery when endoscopic treatment cannot provide safe or oncologically adequate control.

Assess the patient before scheduling and again immediately before sedation. Establish the indication, anticipated interventions, urgency, prior endoscopic and anesthetic experiences, medication list, allergies, bleeding history, cardiopulmonary symptoms, pregnancy possibility, and the patient’s ability to follow preparation instructions. Record baseline blood pressure, heart rate, oxygen saturation, weight, hydration, nutritional status, mental status, and functional capacity. Examine the heart and lungs, identify active infection or decompensation, and assess the abdomen for distension, peritonism, or suspected obstruction. Frailty, sarcopenia, malnutrition, and dependence for activities of daily living predict limited physiologic reserve; they should influence whether an elective procedure is appropriate, whether hospital-based care is safer, and whether the expected benefit justifies the burden. In patients being considered for percutaneous endoscopic gastrostomy , low albumin, systemic inflammation, and very low creatinine, an indirect marker of reduced muscle reserve, identify a particularly high-risk phenotype, although retrospective associations do not establish causation ^[55].

Assign the American Society of Anesthesiologists physical status classification and document the conditions that determine it; the class supports communication but does not replace individualized risk assessment. Ask specifically about exertional chest pain, syncope, orthopnea, heart failure, recent myocardial infarction, significant arrhythmia, pulmonary hypertension, chronic obstructive pulmonary disease, obstructive sleep apnea, home oxygen, and prior difficult ventilation or intubation. Delay elective endoscopy for unstable angina, decompensated heart failure, uncontrolled arrhythmia, severe hypoxemia, or active sepsis until the reversible problem is treated. Obtain targeted investigations rather than routine testing: electrocardiography for relevant cardiac disease or symptoms, hemoglobin and platelet count when bleeding or cytopenia is plausible, electrolytes and creatinine when bowel preparation or medication clearance may be hazardous, and liver tests or coagulation studies when advanced hepatic dysfunction is suspected. Coordinate with anesthesia for severe cardiopulmonary disease, ASA class IV or V, anticipated deep sedation, a difficult airway, or a therapeutic procedure likely to be prolonged.

Inspect mouth opening, dentition, mandibular protrusion, neck mobility, thyromental distance, facial anatomy, and the Mallampati class, the proportion of the oropharynx visible when the patient opens the mouth and protrudes the tongue. Identify obesity, a thick neck, prior difficult airway, stridor, severe sleep apnea, and inability to lie supine. These findings increase the chance that sedation-induced pharyngeal collapse will require airway intervention or conversion to general anesthesia. In obese patients, an oropharyngeal airway that maintains patency and permits capnography reduced hypoxia compared with a standard nasal cannula in a randomized trial, although the study population and device may not generalize to every unit ^[43]. Use supplemental oxygen, continuous pulse oximetry, blood-pressure cycling, electrocardiography when cardiac risk warrants it, and visible end-tidal carbon-dioxide monitoring for deep sedation, high-risk patients, and procedures in which airway rescue may be difficult. In a multicenter device evaluation of older adults, visible capnography was associated with shorter subclinical hypoxemia and more frequent early airway interventions; the observational design supports monitoring but does not prove that capnography caused the reduction ^[42].

Ask about pregnancy in anyone who could be pregnant. Confirm the gestational age and indication, defer nonurgent procedures when feasible, and involve obstetrics and anesthesia when endoscopy is necessary. When delay risks maternal or fetal harm, use the lowest effective sedative exposure, avoid unnecessary fluoroscopy, minimize procedure time, and treat maternal hypoxemia or hypotension promptly because maternal physiologic instability threatens the fetus. Review renal and hepatic function before prescribing bowel-cleansing agents or sedatives: renal impairment increases the danger of volume and electrolyte disturbance and prolongs clearance of some drugs, whereas hepatic dysfunction can increase sensitivity to benzodiazepines and opioids and may coexist with coagulopathy. Prefer polyethylene glycol in patients vulnerable to fluid or electrolyte shifts and individualize the volume and supervision plan. For diabetes, obtain a recent glucose value, prescribe a written insulin and oral-agent plan, reduce or omit prandial insulin while intake is restricted, usually hold sodium-glucose cotransporter-2 inhibitors for 3 days when clinically feasible, and check capillary glucose before preparation, on arrival, and during prolonged fasting. Correct hypoglycemia immediately with oral or intravenous glucose; avoid prolonged fasting without a monitoring plan. A retrospective hospitalized cohort found peri-procedural hypoglycemia in 15.2% of patients undergoing colonoscopy, supporting active rather than passive glucose surveillance ^[50].

Document implanted pacemakers, implantable cardioverter-defibrillators, neurostimulators, insulin pumps, vascular access devices, prosthetic valves, and other hardware. Most diagnostic procedures require no device alteration, but electrosurgical therapy warrants a device-specific plan: identify the device and dependence on pacing, use the lowest effective energy with short applications and a return electrode positioned so current does not cross the generator, and arrange interrogation or magnet/reprogramming advice when clinically indicated. Ask about previous oversedation, hypoxemia, aspiration, difficult mask ventilation or intubation, paradoxical agitation, severe nausea, delayed recovery, malignant hyperthermia, and difficult intravenous access. A prior adverse event changes the anesthetic plan even when the previous record is unavailable; obtain the old anesthetic record when possible.

Consent

Explain the indication, expected benefit, alternatives, including imaging, capsule examination, surveillance, or no procedure, and the consequences of declining. Consent must cover the possibility that the examination will be incomplete because of intolerance, obstruction, poor preparation, altered anatomy, retained gastric contents, or an unsafe clinical condition. Discuss planned and possible interventions: biopsy, polypectomy, mucosal resection, dilation, hemostasis, stenting, foreign-body removal, sphincter therapy, tissue acquisition, and drainage when relevant to the procedure. Explain that an unexpected lesion may require a separate procedure, transfusion, interventional radiology, or surgery; obtain consent for these contingencies rather than assuming that consent for diagnosis includes consent for every therapy.

Describe procedure-specific adverse events in terms the patient can understand. Depending on the examination, these include pain, bleeding, perforation, aspiration, hypoxemia, hypotension, arrhythmia, pancreatitis, cholangitis, infection, adverse drug reaction, dental injury, incomplete treatment, and the need for hospitalization or urgent surgery. For ERCP, discuss post-ERCP pancreatitis, bleeding, infection, perforation, and cardiopulmonary events; dialysis patients have higher procedural vulnerability, and a multicenter retrospective study associated continued aspirin use with post-ERCP bleeding after stone removal ^[54]. Confirm understanding, answer questions, document the chosen sedation strategy, and record who may make decisions if an unanticipated finding requires escalation.

Preparation

For upper endoscopy and procedures involving sedation, use fasting instructions that distinguish solids from clear liquids and follow local anesthesia policy; a common minimum is 6 hours for a light meal and 2 hours for clear liquids, with longer fasting when gastric emptying is delayed or obstruction, severe reflux, gastroparesis, opioid use, or retained food is suspected. Consider gastric ultrasound or anesthesia assessment when aspiration risk is uncertain; evidence for routine gastric-volume assessment before endoscopy remains limited ^[44]. Do not proceed with elective sedation when the patient has violated fasting instructions unless the indication is urgent and the airway plan has been reconsidered.

For colonoscopy, prescribe a split-dose polyethylene glycol regimen: a first portion the evening before and the second portion 4-6 hours before the examination, completed at least 2 hours before sedation according to local policy. Use a low-residue diet on the day before colonoscopy, maintain clear liquids during the final fasting interval, and give written timing, fluid, medication, and rescue instructions. Constipation, diabetes, prior inadequate preparation, opioid use, immobility, hospitalization, and renal dysfunction warrant individualized regimens and early reinforcement. Prediction models identify diabetes and constipation frequently, but their pooled discrimination is only moderate and all models had substantial risk of bias; do not use a score as a substitute for clinical judgment ^[45].

Assess preparation during the examination with the Boston bowel preparation scale (BBPS). A total BBPS of at least 6, with every colonic segment scoring at least 2, generally permits standard surveillance recommendations; randomized-trial meta-analysis found no significant difference in adenoma detection compared with BBPS of at least 8 ^[46]. If preparation is inadequate, irrigate and suction during the procedure when this can restore a safe, complete examination; otherwise document the limitation, avoid false reassurance from a negative result, and repeat colonoscopy with an intensified regimen at an interval determined by the indication and the extent of the inadequacy.

Use antibiotic prophylaxis selectively, not as routine protection for every endoscopic examination. Consider it when incomplete biliary drainage is expected, cholangitis risk is high, severe neutropenia or selected immunosuppression is present, or a specific cardiac or surgical indication exists under current local guidance. For scheduled bile-duct intervention in primary sclerosing cholangitis, a single peri-interventional dose was not associated with a higher rate of post-ERCP cholangitis than a historical multi-day regimen, but the retrospective design and difference in cholangitis severity limit generalization ^[51]. Choose the agent and timing according to the target organisms, allergy history, renal function, and institutional resistance pattern.

Manage antithrombotic treatment by separating procedural bleeding risk from thromboembolic risk. Diagnostic EGD or colonoscopy with mucosal biopsy and low-risk EUS generally permits continuation of aspirin, warfarin with a therapeutic INR, and often P2Y12 inhibitors or direct oral anticoagulants (DOACs), subject to local protocol. Polypectomy, EMR, ESD, dilation, sphincterotomy, ampullectomy, PEG placement, and therapeutic EUS are high-risk procedures. Continue aspirin when its cardiovascular indication is strong, but discuss interruption with the prescribing specialist for exceptionally high-risk resections or when bleeding consequences are severe. Hold clopidogrel, prasugrel, or ticagrelor before high-risk intervention for the recommended platelet-recovery interval, typically 5-7 days, unless recent coronary stenting or acute coronary syndrome makes interruption unsafe. For warfarin, stop sufficiently in advance to achieve the procedural INR target and bridge only when thromboembolic risk is high; resume promptly after hemostasis. Withhold DOACs for a period determined by bleeding risk, renal clearance, and the specific drug, commonly 1 day for low-risk and 2 days for high-risk procedures, and avoid routine heparin bridging. Confirm the plan in writing with cardiology, hematology, or anticoagulation services when risk is high. Dialysis, renal failure, and sphincterotomy require particular caution because both drug exposure and bleeding risk may be increased ^[54].

Sedation and anesthesia

Match the depth of sedation to patient risk, procedure duration, stimulation, position, aspiration risk, and the likelihood of therapeutic intervention. No sedation avoids systemic respiratory and cardiovascular depression and permits immediate recovery, but discomfort and movement can reduce tolerance and procedural quality. Topical pharyngeal lidocaine can facilitate unsedated EGD; in a randomized trial, spray plus viscous lidocaine produced less pain and higher patient satisfaction than spray alone ^[56].

For moderate sedation , give incremental midazolam intravenously, commonly 0.5-2 mg initially and 0.5-1 mg every 2-3 minutes, with lower doses in older adults, frailty, hepatic dysfunction, or respiratory disease. Add fentanyl 25-50 micrograms intravenously in small increments for pain, or meperidine 25-50 mg where local practice supports it; reduce both agents when combined because their respiratory effects are additive. Reversal is available with flumazenil 0.2 mg intravenously, repeated by 0.1-0.2 mg to a usual maximum of 1 mg, and naloxone 0.04 mg intravenously, repeated every 2-3 minutes and titrated to ventilation rather than complete analgesia. Observe for resedation because reversal may wear off before the sedative or opioid.

For monitored anesthesia care , an anesthesia professional administers titrated propofol , often beginning with 10-20 mg boluses or an infusion of approximately 25-75 micrograms/kg/min, aiming for deep sedation while preserving spontaneous ventilation. Propofol has rapid onset and recovery but can cause apnea, airway obstruction, hypotension, and bradycardia; the team must be able to rescue a patient who becomes unresponsive or requires positive-pressure ventilation. Propofol’s rapid recovery is useful, but concern about respiratory and cardiovascular depression explains why its use requires appropriate monitoring and airway expertise ^[39]. Remimazolam, an ultra-short-acting benzodiazepine reversible with flumazenil, is an alternative where approved; a 2026 meta-analysis found faster full alertness and discharge than with midazolam, while safety outcomes were comparable, although only four comparative studies were included ^[40].

Use general anesthesia with endotracheal intubation when airway protection is essential, aspiration risk is high, the procedure is prolonged or highly stimulating, the patient cannot cooperate, or deep sedation is unsafe or repeatedly inadequate. Balance the improved control of ventilation and immobility against hemodynamic effects, airway trauma, longer recovery, and the need for anesthetic resources. A prospective randomized trial in frail adults aged 70-90 years undergoing emergency ERCP found fewer adverse cardiovascular events with conscious sedation than with intravenous general anesthesia, but its single-center population and emergency ERCP context do not justify applying the result to every patient or procedure ^[41].

Regardless of the chosen technique, assign a clinician whose sole responsibility is monitoring when the proceduralist cannot do so, record level of consciousness, respiratory effort, oxygen saturation, blood pressure, heart rate, and capnography when indicated, and have suction, oxygen, bag-mask ventilation, airway adjuncts, defibrillation, and reversal drugs immediately available. Escalate early rather than waiting for persistent desaturation or hypotension: stop the procedure, stimulate the patient, open the airway, provide supplemental oxygen and jaw thrust, suction secretions, use an airway adjunct, and begin assisted ventilation when spontaneous breathing is inadequate. Visible capnography is particularly useful because it can prompt earlier airway intervention before hypoxemia becomes severe ^[42].

Pearl: The safest endoscopy is planned around the intervention that may be required, not merely the examination initially requested, because bleeding risk, aspiration risk, antithrombotic management, fasting, and sedation depth all change when biopsy becomes resection, dilation, sphincterotomy, stenting, or drainage.

A flexible video endoscope is a steerable, articulating instrument that transmits an image from its distal tip to a processor and monitor. Its practical components are the insertion tube, bending section, distal lens and light source, working channel, insufflation and irrigation channels, and control knobs for tip angulation. The working channel must accommodate the intended accessories; suction and irrigation must remain effective when the channel contains fluid or debris. Reusable systems require validated cleaning, disinfection, drying, and storage, whereas single-use systems avoid post-procedure reprocessing, although comparative clinical evidence remains limited. In a prospective two-center study, single-use gastroscopy and colonoscopy achieved predefined-site imaging in all 120 examinations without conversion or device-related adverse events within 24 hours; these findings establish feasibility, not equivalence for all procedures or interventions. ^[2]

Use the highest-resolution system available for the examination. High-definition imaging improves the visibility of subtle surface and vascular changes, but image quality still depends on a clean lens, adequate luminal distension, controlled focus, stable scope position, and sufficient inspection time. Record the examination when local quality assurance permits; recorded video allows objective review of insertion, withdrawal, landmark recognition, and completeness. In one large cohort, artificial-intelligence analysis of recorded colonoscopy correlated strongly with manually measured withdrawal time (r=0.91), but such tools remain adjuncts to endoscopist assessment rather than substitutes for it. ^[60]

Optical enhancement changes how mucosal contrast is displayed. Narrow-band imaging uses selected blue and green wavelengths to accentuate superficial vessels and mucosal architecture; other systems use linked-color imaging, texture and color enhancement, or digital contrast algorithms. Activate enhancement after white-light orientation has established the anatomy, then compare the suspicious area with adjacent mucosa under both modes. Computer-aided quality systems may provide feedback about mucosal exposure or withdrawal speed, but performance can deteriorate when image resolution or acquisition conditions change; one gastrointestinal imaging study found AUROC 0.994 with random-split testing but 0.723 under resolution shift. ^[71] Treat an automated alert as a prompt to inspect, not as a diagnosis.

Chromoendoscopy adds contrast either by spraying a dye or by applying a topical chemical that alters mucosal color. Indigo carmine pools in shallow depressions and accentuates surface relief; a protocol under evaluation for high-risk gastric surveillance uses 20-40 mL of 0.1-0.4% indigo carmine after an initial white-light examination. ^[62] Spraying is most useful after washing retained mucus and reducing bubbles, because dye cannot define a surface that the endoscopist cannot see. Do not allow enhanced imaging to replace a deliberate white-light examination: the evidence for improved lesion detection varies by organ, indication, operator, and technique, and the prospective multicenter gastric study was designed because the incremental benefit of indigo carmine remains uncertain. ^[62]

Insufflation separates mucosal folds and creates a stable field. Use the minimum volume that exposes the lumen, suctioning excess gas during withdrawal and before completion. Air is inexpensive and widely available but is absorbed slowly and may increase post-procedure distension; carbon dioxide is absorbed more rapidly and is often preferred when prolonged examination, difficult decompression, or patient discomfort makes retained gas undesirable. Water can be introduced to clear the lens and float debris; water exchange replaces much of the gas during insertion with water that is suctioned during withdrawal, reducing luminal distension and potentially improving exposure. A retrospective propensity-score-matched study of 4,322 colonoscopies found higher adenoma detection with water exchange than air insufflation (57.9% versus 44.0%) and better preparation scores, but insertion took longer (11.0 versus 6.0 minutes) and total procedure time was longer (22.6 versus 15.6 minutes). ^[57] These observational data support water exchange as an available insertion strategy, not a universal requirement.

A transparent distal cap or other mucosal-exposure attachment can stabilize the tip, maintain distance from the mucosa, flatten folds, and improve visualization behind haustral edges. Select the attachment according to the examination and the operator’s familiarity; it must not obscure the field, impair suction, or prevent passage through a narrowed segment. Evidence for combining mucosal-exposure devices with artificial intelligence is developing: a meta-analysis of 12 studies found higher adenoma detection with combined artificial intelligence and a mucosal-exposure device than with artificial intelligence alone (RR 1.15, 95% CI 1.07-1.25), with the strongest subgroup signal for Endocuff, but no significant difference in several other detection and timing outcomes. ^[68] Irrigate before declaring a segment adequately examined, and suction fluid, bubbles, residual stool, and dye when they obscure the mucosa. Confirm that the suction channel is patent before starting and intermittently during the procedure.

Systematic EGD technique

Begin EGD with the patient, sedation, oxygenation, and scope orientation already stable. Introduce the lubricated endoscope under direct vision, keeping the tip centered and avoiding blind advancement. In the pharynx, identify the hypopharyngeal inlet and enter the esophagus gently; resistance, coughing, or unexpected anatomy warrants withdrawal and reorientation rather than force. Advance through the esophagus while inspecting the lumen, mucosal surface, and gastroesophageal junction. At the junction, document the relationship of the diaphragmatic impression, gastroesophageal junction, and squamocolumnar junction when visible.

Enter the stomach with the lumen centered, then inspect the cardia, fundus, body, incisura, antrum, and pylorus in a reproducible sequence. Aspirate pools that obscure the dependent stomach, insufflate only enough to unfold the rugal surfaces, and rotate the instrument rather than relying on repeated forward movement. Pass the pylorus under direct vision and identify the duodenal bulb, sweep, and second portion. Withdraw slowly through the duodenum and stomach, repeating areas that were poorly distended or obscured. Retroflex in the stomach only after the tip is in a safe, adequately insufflated position; use the maneuver to inspect the cardia, fundus, and proximal lesser curvature, then return to a forward view before withdrawing through the esophagus.

Capture representative images of the major landmarks and every clinically relevant abnormality. For each lesion, record its location using a recognized anatomic segment, its size in millimeters or centimeters, its morphology using a standardized classification when applicable, its surface and color, and the image mode used. Document whether the esophagus, stomach, and duodenum were completely examined, which landmarks were reached, the quality-limiting factors, and any interruption or aborted extent. High-definition scope use and video recording were associated with better adherence to endoscopic quality indicators in a survey, supporting their use as quality infrastructure rather than optional decoration. ^[58]

Systematic colonoscopy technique

Begin insertion with the lumen centered and the tip advanced under direct vision. Minimize looping by reducing redundant insertion, using torque and angulation, and withdrawing when forward pressure no longer advances the tip. Water exchange may be selected when limiting distension or maximizing mucosal exposure is prioritized, accepting its longer insertion time. ^[57] Use suction to decompress the colon and irrigation to clear the lens and residual stool; do not advance through an unrecognized blind space.

Identify the rectum, rectosigmoid junction, sigmoid and descending colon, splenic flexure, transverse colon, hepatic flexure, ascending colon, cecum, and appendiceal orifice. The ileocecal valve and appendiceal orifice provide complementary evidence of cecal intubation; photograph the appendiceal orifice and ileocecal valve, and record the extent reached. Intubate the terminal ileum when the indication requires assessment of the distal small bowel or when the ileocecal valve and cecal landmarks are otherwise uncertain; routine ileal intubation is not required for every asymptomatic screening examination. ^[66][67]

Inspect during withdrawal, not only during insertion. Reduce the insufflation, rotate the shaft, change the tip angle, and expose the proximal sides of folds, flexures, and the underside of haustral ridges. Wash and suction until the mucosa is visible; if a segment remains obscured, repeat the view after repositioning. A deliberate inspection pattern matters because lesion detection is not uniform across the endoscopic field: a prospective observational study found fewer lesions initially seen in the upper-left than the lower-right visual direction (42.4% versus 57.6%). ^[65] Record withdrawal time separately from insertion and therapeutic time; a longer withdrawal is meaningful only when paired with adequate exposure and careful inspection. Computer-aided quality feedback increased withdrawal time by 0.64 minutes in a meta-analysis of randomized trials, illustrating that time is a process measure rather than a standalone guarantee of examination quality. ^[59]

After cecal inspection, withdraw through the ascending, transverse, descending, and sigmoid colon, then inspect the rectum. Retroflexion in the rectum can expose the distal rectal wall and anal transition zone that may be hidden in forward view; perform it only when the lumen is adequately distended and the tip can be controlled safely. If retroflexion is not technically safe or cannot be completed, document the limitation and inspect the distal rectum carefully in forward view. End the examination only after confirming the documented extent, reviewing the recorded images, and stating whether the examination was complete.

Preparation, lesion description, and reporting

Score bowel preparation by segment with the Boston Bowel Preparation Scale (BBPS), assigning each colon segment a score from 0 to 3 after washing and suctioning: 0, unprepared with mucosa not seen; 1, part of the mucosa seen but other areas not adequately visualized; 2, minor residual staining, stool, or opaque liquid but mucosa adequately seen; and 3, entire mucosa seen with no residual staining, stool, or opaque liquid. Report the right, transverse, and left segment scores and their total, rather than reporting only “good” or “poor.” A total BBPS of at least 6 with every segment at least 2 generally permits standard surveillance recommendations; a systematic review found no significant adenoma-detection difference between this intermediate-quality threshold and BBPS of at least 8. ^[46] If any segment remains below 2 after washing and suction, state that the preparation is inadequate for that segment and explain the consequence for examination completeness.

Measure lesion size against an open biopsy forceps, snare, or a calibrated cap when available, and record the estimate before any manipulation changes its shape. Describe morphology with a standardized system such as the Paris classification when appropriate, specifying whether the lesion is pedunculated, sessile, flat, depressed, or has a laterally spreading configuration. Record the segment, distance from a fixed landmark when useful, clock-face position, and relation to folds, valves, scars, or other permanent anatomy. For a lesion that requires future localization, place a tattoo in a documented position relative to the lesion, record the agent and approximate number and location of injections according to local protocol, and photograph both the lesion and the tattoo. Do not rely on the tattoo alone: the report must state whether it is distal or proximal to the lesion and its measured distance.

Use the Minimal Standard Terminology to structure the report so that anatomy, extent, findings, interventions performed, complications, and follow-up recommendations are recorded in consistent terms. The report should identify the indication, extent reached, completeness, preparation quality, sedation, insertion and withdrawal times when relevant, image documentation, lesion descriptors, and limitations. Quality surveys show that documentation is often incomplete: in one survey, 71% of gastroenterologists did not record gastroscopy duration and only 62% used photodocumentation, whereas 90.85% reported using BBPS during colonoscopy. ^[58] A complete report is therefore not an administrative afterthought; it is the permanent record that allows another clinician to judge what was actually examined and what remains uncertain.

Esophagogastroduodenoscopy (EGD) is a mucosal examination whose diagnostic value depends on disciplined inspection rather than rapid passage. Use a high-definition instrument, clean and distend the lumen, inspect the esophagus and gastroesophageal junction, examine the stomach including retroflexion, and inspect the first and second parts of the duodenum. Record landmarks, abnormalities, lesion size and morphology, targeted images, tissue acquisition, and limitations. High-definition endoscopy improves recognition of subtle gastric and duodenal neoplasia, while image-enhanced endoscopy (IEE) helps characterize surface architecture and vascular pattern; neither replaces histopathology when tissue diagnosis is required. ^[77][58]

Mucosal assessment and lesion characterization

In the esophagus, describe reflux injury with the Los Angeles classification , documenting the extent of mucosal breaks rather than using nonspecific terms such as “esophagitis.” Look for circumferential erosions, ulcers, peptic narrowing, and complications such as Barrett esophagus . A salmon-colored mucosal extension above the gastroesophageal junction should be distinguished from an irregular Z-line, gastric folds, and inflammatory change. Measure Barrett mucosa using the Prague classification : the circumferential extent is recorded as C and the maximal proximal extent as M, both in centimeters. Visible nodularity, ulceration, depression, or an irregular vascular or surface pattern warrants targeted sampling and referral for expert assessment because dysplasia or early adenocarcinoma may be focal. ^[58][72]

Inspect for the phenotypic spectrum of eosinophilic esophagitis (EoE), including fixed rings, transient rings, longitudinal furrows, white exudates, edema with loss of vascular markings, mucosal fragility, and strictures. Normal-appearing mucosa does not exclude EoE: exudates and furrows correlate with histologic inflammation, whereas rings and strictures may reflect remodeling. ^[81] Obtain esophageal biopsies when the clinical picture suggests EoE, including when the mucosa appears normal; disease is patchy, and diagnostic sensitivity improves when sampling includes distal and middle esophageal mucosa. ^[75][87] A focal obstructing ring, web, or stricture also requires consideration of esophageal cancer , reflux-related scarring, EoE, caustic injury, medication injury, and postsurgical or radiotherapy change. Do not dilate a suspicious, ulcerated, or irregular narrowing before adequate evaluation for malignancy.

In the stomach, distinguish erythema and edema of gastritis from erosions, hemorrhage, atrophy, intestinal metaplasia, and infiltrative disease. Map the distribution of pale thin mucosa, loss of folds, visible vessels, nodularity, and sharply demarcated intestinal-type mucosa; extensive or corpus-predominant atrophy should prompt histologic assessment for Helicobacter pylori , autoimmune gastritis, and metaplastic risk. Autoimmune gastritis may begin with reddish edematous areae gastricae or longitudinal pseudopolyps before typical corpus-predominant atrophy develops. ^[79] Gastric intestinal metaplasia may be subtle and multifocal, so endoscopic recognition should trigger systematic, site-labeled sampling and risk assessment rather than reliance on a single visible patch. ^[78]

A gastric ulcer is a diagnosis of appearance, not of etiology. Malignancy is suggested by an irregular or heaped-up margin, nodularity, friability, spontaneous bleeding, converging folds that terminate abruptly, or a nonuniform base; benign peptic ulcers more often have smooth regular edges and folds that reach the crater. Sample a suspicious ulcer at the index examination when safe, and regard negative tissue as insufficient reassurance when the morphology remains concerning. Evidence for gastric-ulcer biopsy strategy is limited, heterogeneous, and largely observational; targeted sampling should be intensified for infiltrative or otherwise suspicious lesions. ^[76] Test for Helicobacter pylori when gastritis or peptic ulcer disease is present, because infection is associated with progression across the gastric precancerous sequence. ^[91]

Subepithelial lesions appear as smooth, rounded elevations covered by intact mucosa. A central umbilication favors ectopic pancreas, whereas a firm or ulcerated lesion raises concern for gastrointestinal stromal tumor or another neoplasm; extrinsic compression can mimic either. Use endoscopic ultrasonography to define the wall layer of origin, echogenicity, borders, and relation to adjacent structures, and obtain tissue when the result will change management. Histopathology remains necessary for definitive diagnosis when ectopic pancreas or another subepithelial mimic cannot be recognized confidently by its endoscopic appearance. ^[84]

Duodenal inspection must include the bulb and second portion because superficial non-ampullary duodenal epithelial tumors may be flat, slightly elevated, depressed, granular, or discolored. Under white light, document the lesion’s extent and relationship to the ampulla; use IEE to assess irregular surface and vascular patterns. Suspicious lesions require expert characterization and tissue diagnosis, while management should account for the duodenum’s substantial risk of treatment-related complications. ^[77] A large pedunculated or submucosal-appearing lesion may instead represent Brunner gland hyperplasia, lipoma, ectopic pancreas, or another benign process; a depressed, red, ulcerated, or non-lifting component should prevent dismissal as benign. ^[85]

Upper gastrointestinal bleeding

During EGD for upper gastrointestinal bleeding , first identify the bleeding source and assess whether blood obscures additional lesions. Describe active spurting or oozing, an oozing lesion without a visible vessel, a nonbleeding visible vessel, an adherent clot, and a clean or pigmented ulcer base using the Forrest classification . Forrest Ia denotes spurting bleeding, Ib oozing bleeding, IIa a nonbleeding visible vessel, IIb an adherent clot, IIc a flat pigmented spot, and III a clean base. The classification communicates rebleeding risk and determines whether endoscopic hemostasis is indicated: active bleeding and a nonbleeding visible vessel require endoscopic therapy, an adherent clot requires careful consideration of clot removal and treatment of the underlying stigma, whereas a flat pigmented spot or clean base generally does not require endoscopic hemostasis. ^[76]

Record the lesion’s location, size, base, surrounding mucosa, and suspected cause. In a patient with portal hypertension, distinguish esophageal or gastric varices from folds, prominent submucosal vessels, and extrinsic compression; also inspect for portal hypertensive gastropathy , characterized by a mosaic or “snake-skin” pattern with superimposed red signs. Varices, portal hypertensive gastropathy, peptic ulcer disease, malignancy, and vascular ectasia require different treatment pathways, so the endoscopic report should identify the presumed source rather than merely record “blood in stomach.” Noninvasive criteria can reduce screening EGD in selected compensated cirrhosis populations, but EGD remains the reference examination when high-risk varices are suspected or when clinical circumstances require direct assessment. ^[83][89]

Superficial neoplasia and infiltrative disease

Use the Paris classification for superficial epithelial neoplasia: type 0-I is protruding, type 0-II is superficial and subdivided into slightly elevated (0-IIa), flat (0-IIb), and slightly depressed (0-IIc), and type 0-III is excavated. Describe color, surface pattern, ulceration, spontaneous friability, borders, and non-lifting behavior in addition to the Paris type. These features help distinguish a lesion suitable for endoscopic resection from one requiring staging or surgery, but invasion depth and lymphovascular risk are histopathologic determinations. ^[58][78]

With high-definition white light, establish anatomy and lesion boundaries before switching to IEE. Narrow-band imaging (NBI) accentuates superficial vessels and mucosal architecture; blue-light or comparable virtual chromoendoscopy can provide similar contrast where available. Dye-based chromoendoscopy, particularly with indigo carmine, pools in depressions and sharpens lesion borders; wash away mucus and bubbles first. Use enhancement to target biopsies and delineate visible lesions, not to replace a complete white-light examination or histopathology. AI systems may assist lesion detection or characterization, but current upper-GI evidence is dominated by retrospective studies using selected images and has limited external, prospective, real-world validation. ^[82]

Malignant or infiltrative disease may present as an ulcerated mass, irregular depressed lesion, nodular mucosa, loss of distensibility, thickened folds, a narrowed lumen, or apparently nonspecific erythema. Linitis plastica and other infiltrative tumors can yield superficial negative biopsies because the neoplastic process lies beneath relatively preserved epithelium. When the appearance and pathology disagree, repeat targeted sampling, cross-sectional imaging, and endoscopic ultrasonography -guided tissue acquisition are appropriate; do not label the lesion benign solely because superficial biopsies are negative. ^[76][84]

Pearl: A normal-looking surface does not exclude EoE, infiltrative cancer, or early neoplasia; when the clinical picture and endoscopic appearance disagree, obtain adequate tissue and escalate to EUS or cross-sectional imaging rather than accepting a reassuring visual impression. ^[75][81][82]

Colonoscopy is the definitive lower-gastrointestinal examination when the clinical question concerns mucosal disease, colorectal neoplasia, bleeding, or tissue diagnosis. It is used for colorectal cancer screening, evaluation of a positive fecal immunochemical test (FIT), hematochezia, iron-deficiency anemia, chronic diarrhea, suspected inflammatory bowel disease (IBD), and surveillance after colorectal neoplasia. Colonoscopy remains the most comprehensive screening modality because it can detect and remove precursor lesions, although its invasiveness, cost, logistical demands, and lower participation limit its use as the universal first-line population test. ^[107] A positive FIT requires colonoscopy even when symptoms are absent; after a high-quality examination that finds no explanatory lesion, routine upper-gastrointestinal or small-bowel evaluation and routine early repeat colonoscopy are not supported for otherwise asymptomatic patients, although residual colorectal cancer risk is low rather than zero. ^[99]

Examine the perianal skin and anus before intubation, then inspect the rectum, including the dentate line, anal canal, and retroflexed view when safe, before withdrawing through the sigmoid, descending, transverse, and right colon. Confirm cecal intubation with the appendiceal orifice and ileocecal valve; intubate the terminal ileum when the indication or abnormal ileocecal appearance makes proximal small-bowel assessment useful. During withdrawal, clean residual fluid and debris, flatten folds with controlled insufflation or water, and inspect the proximal and distal aspects of each fold rather than relying on a central forward view. A negative examination is only as reliable as its completeness and mucosal exposure; missed lesions remain a clinically relevant limitation of colonoscopy. ^[100]

Describe every lesion by location, size, surface, color, border, and relationship to folds or landmarks. The Paris classification provides the gross morphology of superficial colorectal lesions: 0-I lesions are protruding, 0-II lesions are nonprotruding and subdivided into slightly elevated (0-IIa), completely flat (0-IIb), or slightly depressed (0-IIc), and 0-III lesions are ulcerated. Depression, ulceration, marked stiffness, fold convergence, spontaneous bleeding, or loss of surface pattern increases concern for submucosal invasion and should prompt expert characterization rather than routine removal. Histopathology, not optical impression alone, determines invasion depth, margin status, and lymphovascular risk.

For diminutive or small lesions, narrow-band imaging (NBI) optical assessment can supplement white-light examination. The NICE classification uses color, vessels, and surface pattern: type 1 lesions are usually non-neoplastic, type 2 lesions usually represent adenoma or superficial neoplasia, and type 3 lesions suggest deep submucosal invasion. The JNET classification refines NBI interpretation: type 1 favors hyperplastic polyp or sessile serrated lesion without dysplasia; type 2A favors low-grade conventional adenoma; type 2B raises concern for high-grade dysplasia or shallow invasion; and type 3 suggests deep invasion. Optical classification guides whether a lesion can be managed locally, requires en bloc assessment, or should be referred for oncologic evaluation, but it does not replace tissue diagnosis when invasion or dysplasia is suspected. Computer-aided detection may increase detection, yet pooled diagnostic performance and human interpretation remain imperfect; in a meta-analysis, pooled sensitivity and specificity for colorectal polyps of any size were 0.94 and 0.91, respectively. ^[94]

Distinguish the major epithelial lesion phenotypes. Conventional adenomas are often reddish, protruding, and more sharply demarcated than surrounding mucosa; their dysplastic surface commonly shows regular but crowded or tubular vessels and a structured pattern. Sessile serrated lesions are usually subtle, broad-based, and pale, with indistinct borders, mucus or a “cap,” and an expanded crypt pattern under enhanced imaging; they are particularly easy to overlook behind folds. Hyperplastic polyps are commonly small, pale, and distal, with a regular surface and no dysplastic optical pattern. Optical overlap is substantial, especially between sessile serrated lesions and hyperplastic polyps, so uncertainty should favor retrieval for histologic classification rather than confident visual labeling. Human-machine studies show a tendency to overcall lesions as sessile serrated lesions after rejecting computer-aided predictions, a bias that may increase unnecessary surveillance. ^[92]

In inflammatory bowel disease , inspect the entire colon for continuous or patchy inflammation, architectural distortion, strictures, pseudopolyps, ulcers, and discrete dysplastic lesions. Ulcerative colitis activity should be recorded with the Mayo endoscopic score : 0, normal or inactive mucosa; 1, mild erythema, decreased vascular pattern, and mild friability; 2, marked erythema, absent vascular pattern, friability, and erosions; and 3, spontaneous bleeding and ulceration. The ulcerative colitis endoscopic index of severity (UCEIS) separately grades vascular pattern, bleeding, and erosions or ulcers, producing a score from 0 to 8; recording the component descriptors prevents a single global score from obscuring the dominant lesion. Targeted biopsies from abnormal mucosa and segment-labeled biopsies from apparently normal mucosa are required when the diagnosis, extent, activity, infection, microscopic inflammation, or dysplasia will alter management. Chronic intestinal inflammation creates a dysplasia-prone mucosal field, and IBD-associated dysplasia may be multifocal and biologically distinct from sporadic adenoma. ^[93]

When colitis-associated dysplasia is visible, characterize its border with high-definition white light and enhanced imaging, document whether it is resectable, and obtain targeted biopsies if the diagnosis or extent is uncertain. Poorly demarcated, nonlifting, ulcerated, or multifocal lesions require expert review and often cross-sectional imaging or surgical discussion rather than repeated indiscriminate biopsy. Crystal-violet-based delineation has been reported in a four-case series to improve recognition of lesion extent before endoscopic submucosal dissection, but this remains an adjunctive technique requiring prospective validation. ^[97]

The principal endoscopic patterns encountered in diagnostic colonoscopy are summarized below.

A surveillance examination should be adapted to the prior lesion or disease field rather than treated as a routine screening colonoscopy. In older adults, continuation of colorectal cancer screening or polyp surveillance after age 75 should be individualized according to prior findings, comorbidity, life expectancy, procedural risk, and whether the patient could undergo endoscopic, surgical, or oncologic treatment if significant disease were found. ^[108]

Video capsule endoscopy (VCE) is the preferred mucosal survey when disease is suspected beyond the reach of standard upper and lower endoscopy, particularly after those examinations do not explain bleeding. Suspected small-bowel bleeding (SSBB) includes overt bleeding, melena or hematochezia without an identified source, and occult bleeding, usually unexplained iron-deficiency anemia (IDA). VCE is generally performed first because it is less invasive and can inspect the entire small-bowel mucosa, whereas device-assisted enteroscopy (DAE) is reserved for tissue acquisition or treatment of a lesion identified or strongly suggested by VCE or cross-sectional imaging. ^[109][114]

In unexplained IDA, capsule examination can reveal a small-bowel lesion even when upper and lower endoscopy are unrevealing. In a prospective multicenter study of 170 patients with IDA, VCE identified a lesion with bleeding potential in 50.0% of patients, although lesions judged likely to be causal were found in 10.0%; the finding changed management in 52.3% after pan-gastrointestinal evaluation. ^[13] In overt SSBB, arrange DAE promptly when the patient is actively bleeding and a treatable source is likely: a retrospective cohort found diagnostic yields of 61% for occult, 67% for inactive overt, and 95% for active overt bleeding, with an overall therapeutic yield of 51%; examination within 72 hours of an overt bleed was associated with higher diagnostic yield. ^[114] These retrospective data support early intervention but do not establish an optimal timing threshold for every patient.

VCE is also useful in suspected or established Crohn disease when small-bowel mucosal activity, extent, or treatment response remains uncertain. Capsule findings can expose superficial inflammatory disease that cross-sectional imaging may miss and can support treat-to-target decisions; a 2026 systematic review of nine studies involving 321 patients found active disease in 76.3% at baseline, treatment modification in 71%, and capsule-guided mucosal healing in 27% after follow-up, with substantial heterogeneity and only one episode of retention. ^[111] Use VCE selectively for mucosal surveillance in Crohn disease rather than as a substitute for imaging when obstruction, a stricture, transmural inflammation, fistula, abscess, or an extraluminal mass is suspected. In children with Crohn disease, VCE findings prompted treatment initiation or escalation in 52% of patients assessed for active disease, but one retained capsule revealed previously unsuspected severe stricturing disease. ^[122]

Other appropriate indications include surveillance for small-bowel polyps in hereditary polyposis syndromes, especially Peutz-Jeghers syndrome , assessment of suspected small-bowel tumors when a mucosal lesion is the target, and selected surveillance of mucosal inflammatory burden after treatment. In Peutz-Jeghers syndrome, VCE maps polyp burden and helps plan therapeutic enteroscopy; DAE can then remove selected lesions, including 5-9-mm polyps treated by cold-snare polypectomy in a prospective study. ^[109][113] Capsule surveillance should be tied to a management decision, such as selecting an enteroscopic route, scheduling polypectomy, or assessing treatment response, because VCE cannot biopsy, tattoo, or treat.

Capsule safety and retention

Do not swallow a diagnostic capsule when a fixed or functionally important luminal narrowing is likely. The principal hazards are a known or suspected small-bowel stricture, obstructive symptoms, a suspected obstructing tumor, and Crohn disease with stenotic or penetrating behavior; diaphragm disease from nonsteroidal anti-inflammatory drugs (NSAIDs) is another recognized cause. Capsule retention is defined as persistence of the capsule in the gastrointestinal tract for at least 2 weeks and occurs in approximately 2% of examinations, with Crohn disease, obstructive tumors, and NSAID-associated diaphragm disease among common causes. ^[110] Prior abdominal surgery is not an automatic contraindication, but it raises concern for anastomotic narrowing, adhesional obstruction, altered transit, or an excluded segment and therefore warrants review of operative anatomy and recent imaging. In a matched cohort, patients with surgically altered anatomy had similar completion and diagnostic yields to patients with native anatomy, although patency testing was used more often. ^[115]

When patency is uncertain, give a dissolvable patency capsule before the diagnostic capsule. Its purpose is to demonstrate passage through the bowel or safe dissolution rather than to provide mucosal imaging. Do not proceed to VCE if the patency capsule fails to pass or remains in the small bowel; first localize it and clarify the cause of delayed transit with cross-sectional imaging or enteroscopy. If excretion cannot be confirmed, abdominal ultrasonography can be a radiation-free first assessment: in a prospective study, it identified all six patients with small-bowel patency-capsule retention shown by computed tomography, although nonvisualized findings require cautious interpretation. ^[121] A negative patency test reduces, but does not abolish, retention risk; document that residual risk during consent.

A retained capsule requires localization and a plan based on symptoms, the cause of retention, and accessibility. Use DAE to retrieve a capsule or patency-capsule shell when the retained object is reachable, while persistent obstruction, perforation, tumor, or an inaccessible capsule may require surgical retrieval. DAE also permits treatment of the lesion responsible for retention, such as selected short benign strictures by balloon dilation, although dilation must follow assessment of inflammatory and malignant risk. ^[109][125]

Device-assisted enteroscopy

Push enteroscopy uses a longer endoscope to examine the proximal jejunum and is useful when VCE or bleeding pattern localizes a lesion near the duodenojejunal flexure. It permits direct visualization, biopsy, injection, clipping, thermal hemostasis, and removal of selected lesions within reach, but it cannot reliably survey or treat the distal small bowel. DAE extends this capability by pleating the bowel over the instrument with an overtube and either a balloon system or a spiral mechanism; the procedure is technically demanding and is concentrated in specialized centers. ^[109]

Single-balloon enteroscopy (SBE) and double-balloon enteroscopy (DBE) use an overtube-mounted balloon system to shorten and stabilize the small bowel. DBE uses balloons on both the enteroscope and overtube, whereas SBE uses one overtube balloon with tip angulation and withdrawal maneuvers to achieve bowel shortening. Both provide deep intubation, targeted biopsy, tattooing, polypectomy, hemostasis, dilation of selected benign strictures, and retrieval of foreign bodies or retained capsules; available evidence has not shown a significant overall complication difference between them. ^[109][114] Spiral enteroscopy advances by rotating a spiral overtube to pleat bowel onto the instrument and is another deep-enteroscopy option when local expertise and equipment are available. ^[109]

Choose the oral or anal route from the suspected lesion’s location rather than from the presence of a positive capsule alone. In one BAE cohort, an oral approach was selected when the lesion was before approximately 70% of total small-bowel transit time on VCE; melena generally favored oral assessment, whereas maroon stool or hematochezia favored the anal route when localization was unavailable. ^[114] If complete treatment requires both directions, stage antegrade and retrograde examinations and document the deepest point reached. Tattoo the lesion or the furthest point of insertion when surgery, repeat enteroscopy, or a second route may be needed; tattooing is particularly useful for tumor localization and operative planning. ^[118]

Once the target is reached, clean and distend the segment sufficiently to define its borders, obtain biopsies from suspicious mucosal or mass lesions, and select therapy according to lesion type. Treat bleeding angioectasias commonly with argon plasma coagulation; clips are useful when a focal vessel or Dieulafoy lesion can be captured, and injection therapy is generally an adjunct rather than definitive treatment. In a BAE series of SSBB, vascular lesions accounted for 42% of positive examinations, argon plasma coagulation was used in 75% of treated cases, and hemoclips in 51%. ^[114] For polyps, remove only lesions that are technically accessible and appropriate for endoscopic resection; larger, broad-based, deeply ulcerated, or suspicious lesions require careful staging and multidisciplinary planning. Tumors may require biopsy and tattooing but often still require surgery because enteroscopy does not provide lymph-node clearance or complete oncologic resection. ^[109][118]

Characteristic findings and modality selection

The capsule is best viewed as a detector and map; DAE is the instrument that converts a positive finding into histology or treatment. Common capsule and enteroscopic findings in SSBB include angioectasias, ulcers and erosions, strictures, inflammatory mucosa, polyps, tumors, and less commonly small-bowel varices or diverticular bleeding. In a large SBE series, angioectasia was the dominant vascular lesion, while ulcers were the most frequent nonvascular lesion; malignancy and inflammatory bowel disease were less frequent but clinically consequential. ^[114][117] In Crohn disease, edema and ulceration are typical mucosal abnormalities, while NSAID enteropathy may produce erosions, ulcers, or diaphragm strictures that both explain anemia and increase retention risk. ^[110][123] Small-bowel varices should prompt assessment for portal-hypertensive or venous outflow disease before biopsy or thermal therapy, because the apparent mucosal lesion may represent a high-flow venous channel. ^[117]

Pearl: Use VCE to survey and localize; use patency testing when passage is uncertain; use route-directed DAE when the result must lead immediately to biopsy, hemostasis, polypectomy, dilation, tattooing, or retrieval. ^[109][114]

Endoscopic ultrasonography (EUS) combines a high-frequency ultrasound transducer with endoscopic access, allowing the operator to examine the gastrointestinal wall and immediately adjacent organs from close range. Its diagnostic advantage is resolution: fluid, ductal structures, vascular interfaces, and small regional lesions can be characterized before tissue is obtained. EUS is therefore most useful when the clinical question concerns depth, origin, regional extension, or a lesion that is inaccessible to routine mucosal biopsy.

Examination modes and image interpretation

Radial EUS produces a 360-degree, cross-sectional image perpendicular to the endoscope. It is well suited to mapping the full circumference of the esophagus, stomach, rectum, and mediastinum, defining the relationship of a lesion to the bowel wall, and assessing circumferential tumor spread. Linear, or curvilinear, EUS produces a longitudinal image parallel to the endoscope and maintains a real-time needle path in the imaging plane. Use linear EUS when tissue acquisition, injection, drainage, or vascular intervention is anticipated; use radial EUS when circumferential anatomy and wall-layer mapping are the primary questions.

The normal gastrointestinal wall is recognized as alternating echogenic bands rather than as isolated histologic structures. From the lumen outward, the first bright band represents the superficial mucosa, the second dark band the deeper mucosa or muscularis mucosae, the third bright band the submucosa, the fourth dark band the muscularis propria, and the outer bright interface the serosa or adventitia. A lesion’s layer of origin narrows the differential diagnosis: submucosal lesions are often lipoma, ectopic pancreas, vascular lesions, or cysts, whereas a hypoechoic mass from the fourth layer raises gastrointestinal stromal tumor or leiomyoma. Layer disruption, irregular borders, ulceration, or invasion beyond the muscularis propria increases concern for malignancy and changes management from observation or local excision to tissue diagnosis and oncologic staging.

Describe every lesion by layer of origin, echogenicity, homogeneity, border, size, shape, internal cystic spaces, calcification, ulceration, and relationship to vessels and ducts. EUS does not establish histology from appearance alone: an apparent gastrointestinal stromal tumor can be another mesenchymal or neural neoplasm, and an apparently benign subepithelial lesion may be malignant. In colorectal subepithelial lesions, location, echogenicity, and wall layer were associated with mismatch between EUS and final histopathology; the retrospective study found mismatch in 26.2% of 301 lesions, supporting tissue acquisition when the result will alter treatment. ^[133]

Assess regional lymph nodes in at least two planes. Suspicious nodes tend to be round rather than ovoid, relatively homogeneous and hypoechoic, sharply demarcated, and larger than adjacent nodes, but no single feature reliably distinguishes metastatic from inflammatory disease. Sample the most suspicious accessible node rather than the primary lesion when that will establish malignancy or stage disease with less risk. A negative needle sample does not exclude micrometastatic disease, particularly when the node is small or technically difficult to target.

Use color or power Doppler ultrasound before puncture, injection, or drainage. Doppler identifies arteries, veins, collaterals, varices, and intralesional flow, preventing traversal of a vessel and helping distinguish a vascular structure from a solid mass. A cystic-appearing lesion with internal flow is not a simple cyst; it may represent a pseudoaneurysm, varix, or hypervascular tumor and should not be punctured until vascular anatomy is clarified.

Elastography estimates tissue stiffness by measuring deformation or shear-wave propagation. A stiff, heterogeneous area within a pancreatic mass or lymph node can support targeting, but elastography is an adjunct rather than a histologic diagnosis because inflammation, fibrosis, necrosis, and technical compression can produce overlapping measurements. Contrast-enhanced EUS uses intravenous microbubble contrast to examine enhancement during the arterial and venous phases. Hyperenhancing septa or mural nodules within a cyst, and irregular or hypoenhancing regions within a solid mass, should direct sampling or surgical review; enhancement must be interpreted with morphology and cross-sectional imaging rather than used as a stand-alone cancer test.

Lesion-specific applications

Pancreatic EUS is particularly valuable for a small or isoattenuating pancreatic mass, an indeterminate cyst, suspected chronic pancreatitis, or a lesion requiring tissue before chemotherapy or surgery. Examine the entire gland from the stomach and duodenum, trace the main pancreatic duct, document duct caliber and upstream atrophy, and inspect the common bile duct, portal confluence, splenic vessels, and regional nodes. A pancreatic ductal adenocarcinoma is commonly an irregular hypoechoic mass with ductal obstruction or upstream dilation, but focal pancreatitis can mimic it; obtain tissue when a malignant diagnosis will determine treatment and imaging is not definitive. Structured EUS variables have supported retrospective machine-learning models for pancreatic ductal adenocarcinoma, but these models remain observational and require external clinical validation before replacing expert interpretation. ^[137]

For a pancreatic cyst, document communication with the main duct, size and location, wall thickness, septations, mural nodules, solid components, debris, calcification, and duct dilation. Aspirate cyst fluid selectively when the diagnosis or management is uncertain. Send fluid for cytology, carcinoembryonic antigen, glucose where available, and amylase or lipase when duct communication or a pseudocyst is in question. Thick mucin, an enhancing mural nodule, or a solid component raises concern for a mucinous cystic neoplasm or intraductal papillary mucinous neoplasm; a simple anechoic collection in the appropriate clinical context favors a pseudocyst, but infection and hemorrhage can make fluid complex. Cytology is often limited by scant cellularity, so fluid findings must be integrated with EUS morphology and CT or MRI.

In chronic pancreatitis , EUS may show hyperechoic strands and foci, lobularity, cysts, calcifications, duct irregularity, duct stones, and changes in the duct wall. The combination and distribution of findings matter more than one minor abnormality: fibrosis and calcification can create a heterogeneous gland and obscure a coexisting carcinoma. A new focal mass, abrupt duct cutoff, progressive jaundice, or unexplained weight loss warrants repeat targeted imaging and tissue acquisition even when established chronic pancreatitis is present.

For bile-duct disease, EUS detects distal duct stones, sludge, ampullary lesions, duct-wall thickening, and malignant obstruction, and can sample a pancreatic, distal biliary, or nodal target. Use Doppler to define the portal vein and peribiliary vessels before puncture. EUS-guided biliary drainage is an alternative after unsuccessful or unsuitable ERCP, including transmural hepaticogastrostomy or choledochoduodenostomy. In a retrospective series of EUS-guided hepaticogastrostomy, early adverse events occurred in 15% of patients, including peritonitis in 7%; a punctured bile duct measuring 3 mm or less independently increased peritonitis risk. ^[26] Choose a sufficiently dilated duct and a stable transgastric or transduodenal tract, and perform these procedures in experienced multidisciplinary units.

EUS evaluates gallbladder wall thickening, polyps, stones, sludge, adenomyomatosis, and suspected invasion by gallbladder carcinoma. A small, smooth, pedunculated lesion without concerning vascularity may be a cholesterol polyp; sessile morphology, irregular thickening, loss of wall layers, or an adjacent mass requires cross-sectional imaging and surgical review. For acute cholecystitis in patients unsuitable for immediate surgery, EUS-guided gallbladder drainage with a lumen-apposing metal stent (LAMS) creates a controlled gallbladder-duodenal or gallbladder-gastric fistula. In a multicenter cohort, technical success was 98.2% and clinical success 93.5%; LAMS-related adverse events occurred in 11.1%. These retrospective results support feasibility but do not establish superiority over surgery or alternative drainage. ^[25]

Mediastinal EUS and EUS-guided tissue acquisition can assess paraesophageal, subcarinal, posterior mediastinal, and left-sided hilar lesions when an endoluminal route provides a safe path. Sample a node or mass that will change diagnosis or stage, and avoid puncturing a vascular lesion or traversing an uncorrected high-risk structure. Perirectal EUS defines the relationship of a mass to the rectal wall, mesorectal fascia, prostate or vagina, and pelvic nodes. For rectal cancer, EUS is most useful for distinguishing mucosal or submucosal disease from invasion into the muscularis propria in selected early tumors; pelvic MRI is generally needed for mesorectal fascia, extramural vascular invasion, and broader surgical planning. In rectal deep endometriosis, retrospective data found that EUS and MRI provide complementary rather than interchangeable information. ^[131]

EUS characterizes gastrointestinal subepithelial tumors by layer and internal architecture. A homogeneous hyperechoic lesion in the submucosa favors lipoma; a heterogeneous lesion with central umbilication in the second or third layer favors ectopic pancreas; a hypoechoic fourth-layer lesion suggests gastrointestinal stromal tumor or leiomyoma; and a cystic lesion with posterior enhancement suggests a duplication cyst or other enteric cyst. Obtain tissue when the lesion is symptomatic, enlarging, ulcerated, large, irregular, or likely to require systemic therapy. A 19-gauge core needle, Franseen-tip needle, or other fine-needle biopsy device is preferred when architecture and immunohistochemistry are needed. In a retrospective series of subepithelial lesions, a heparinized wet-suction, single-pass technique achieved definitive histologic diagnosis in 86.1% overall and 95.8% in lesions at least 20 mm, although the study was single-center and lacked a comparative randomized design. ^[135]

Locoregional cancer staging

For esophageal and gastric cancer, EUS assesses depth of mural invasion and regional nodes; for rectal cancer, it assesses mural penetration and perirectal nodes; and for pancreaticobiliary cancer, it assesses vascular contact, ductal obstruction, adjacent-organ invasion, and regional nodes. Use CT or MRI to evaluate distant metastases and extensive vascular or extraluminal disease, because EUS has a limited field and cannot substitute for whole-body staging. Sample a suspicious node or metastatic focus when it will change the treatment pathway. In a potentially resectable pancreatic mass, avoid transgressing a planned surgical field when tissue is unnecessary, but obtain EUS-guided tissue before neoadjuvant or systemic therapy.

EUS-guided tissue acquisition and intervention

Fine-needle aspiration (FNA) uses a slender needle to obtain cells and small tissue fragments, whereas fine-needle biopsy (FNB) uses a core-design needle to preserve tissue architecture. FNA remains useful for cytology, cyst-fluid sampling, and some lymph-node targets; FNB is preferable for lymphoma, mesenchymal tumors, autoimmune disease, and lesions requiring immunohistochemistry or molecular testing. Coordinate the target and specimen plan with pathology before puncture: state the differential diagnosis, request cell block or core processing when required, and reserve material for flow cytometry or molecular studies when clinically indicated. Repeat passes only when adequacy is uncertain or the result will alter management; rapid on-site evaluation, when available, can reduce nondiagnostic sampling, but a core specimen may be more valuable than additional low-cellularity aspirates.

For solid lesions, use Doppler to exclude vessels, puncture through the shortest safe route, and sample viable solid tissue rather than necrotic centers. For cysts, avoid unnecessary puncture, aspirate only when the result changes management, and send fluid in separate labeled containers for cytology and biochemical analysis. A nondiagnostic sample is not reassurance when imaging and clinical findings remain concerning; reassess the target, needle type, route, and need for repeat EUS, alternative biopsy, or surgery.

Selected therapeutic EUS procedures include drainage of pancreatic fluid collections, EUS-guided biliary drainage, gallbladder drainage, celiac plexus block or neurolysis for selected pancreatic-cancer pain, and deployment of LAMS to create a controlled transluminal tract. Drain a mature, encapsulated collection only when infection, persistent symptoms, obstruction, or complications justify intervention; define vascular anatomy and adjacent organs before puncture. EUS-guided celiac plexus neurolysis is palliative and should be discussed as an analgesic intervention, not tumor treatment. LAMS is powerful because its flanges appose two luminal or cavity walls, but migration, bleeding, buried stent, obstruction, and delayed closure remain relevant limitations; use it in experienced centers with a clear plan for follow-up and removal or exchange when required.

ERCP is primarily a therapeutic procedure. Use it when the anticipated benefit is ductal intervention, biliary drainage, stone extraction, sphincter therapy, stricture dilation, tissue acquisition, or stent placement, not simply to establish a diagnosis. For suspected obstruction without an immediate plan for therapy, start with MRCP , contrast-enhanced CT, or EUS ; in cholestatic liver disease, ERCP is reserved for therapeutic or equivocal scenarios ^[159]. In patients with a high likelihood of choledocholithiasis but negative initial imaging, prospective data found that EUS had 100% sensitivity and a 100% negative predictive value, and avoided unnecessary ERCP in 79.3% of patients without stones ^[148].

Indications and procedural objectives

Perform ERCP for confirmed or strongly suspected common bile duct stones when extraction is required; for acute cholangitis when infected bile requires endoscopic decompression; and for symptomatic or clinically consequential benign or malignant biliary obstruction. In acute cholangitis, timing should reflect physiology rather than a rigid clock: a retrospective cohort found the highest mortality among patients undergoing ERCP within 24 hours, but the authors attributed this principally to greater baseline instability, not harm from early drainage ^[147]. Resuscitate, give antimicrobial therapy, correct reversible coagulopathy, and drain urgently when obstruction and infection persist or the patient is deteriorating.

Use ERCP to palliate or prepare treatment for malignant distal or hilar obstruction, to sample an indeterminate stricture, and to restore bile flow when jaundice, pruritus, cholangitis, or impaired oncologic treatment results from obstruction. In a multicenter real-world study of malignant biliary obstruction, risk-adjusted technical success was 91.3%, post-ERCP pancreatitis occurred in 5.4%, additional ERCP was required in 23.7%, and 30-day mortality was 4.2% ^[151]. These figures are benchmarks from observational practice, not guarantees for an individual patient. In a single-center retrospective cohort of suspected malignant or indeterminate strictures, clinical drainage success was 93.3%, but stent-type comparisons were confounded by baseline treatment selection ^[145].

Treat postoperative bile leaks after cholecystectomy, hepatectomy, or liver transplantation when contrast extravasation identifies a leak that can be decompressed by transpapillary drainage. The usual objective is to lower the transpapillary pressure gradient so bile preferentially flows into the duodenum rather than through the defect. A large tertiary-center series found initial clinical success in 80.7% and final success in 96%; high-grade and post-hepatectomy leaks had lower initial success and more often required repeat ERCP ^[146]. A 10-Fr plastic stent was associated with lower odds of delayed closure in one small transplant cohort, but grade-based stent strategies remain unvalidated ^[160].

Selected pancreatic indications include pancreatic sphincter therapy, drainage of pancreatic duct disruption or leak, dilation of dominant benign strictures, extraction or fragmentation of pancreatic duct stones, and stenting for symptomatic obstruction in chronic pancreatitis. Use pancreatography only when the result will guide treatment, because contrast injection and repeated instrumentation of the pancreatic orifice increase procedural burden. Pancreatoscopy can assist assessment of postoperative duct anatomy or intraductal papillary mucinous neoplasia in carefully selected cases; a small case series reported successful target-duct insertion in all 12 procedures using an ultraslim catheter-type cholangioscope/pancreatoscope ^[150].

Technique

Advance the side-viewing duodenoscope to the second portion of the duodenum, identify the major papilla, and orient the elevator and catheter so that the axis of the instrument follows the expected common channel. Confirm the papilla before injecting contrast. Selective biliary cannulation is best performed with a sphincterotome and guidewire under fluoroscopic control, maintaining wire access while exchanging catheters or accessories. Limit contrast injection until the intended duct has been entered; document duct caliber, filling defects, strictures, leak sites, intrahepatic distribution, and drainage.

After biliary access, perform a pancreatogram only for a pancreatic indication or when pancreatic duct anatomy must be clarified. A biliary sphincterotomy opens the papillary outlet for extraction, dilation, and stent delivery. For stones, sweep with an extraction balloon or capture them with a basket after sphincter therapy. When a stone exceeds the papillary opening or is impacted, combine sphincterotomy with controlled large-balloon dilation or use mechanical lithotripsy. A 2026 meta-analysis of 17 randomized trials suggested that sphincterotomy plus large-balloon dilation increased first-session clearance and reduced mechanical lithotripsy and bleeding compared with sphincterotomy alone, but the certainty was low to very low and the authors judged the findings hypothesis-generating rather than practice-changing ^[144].

Balloon dilation must be sized to the distal duct and performed gradually; do not force an extraction device through a fixed or inadequately opened outlet. For small stones, either balloon or basket extraction is reasonable. In a randomized trial of stones measuring 8 mm or less, a spiral basket improved early procedural success and reduced sweeps and procedure time, but did not improve final clearance and was not compared with balloon extraction ^[156]. For large, impacted, recurrent, or anatomically difficult stones, single-operator cholangioscopy permits direct visualization, targeted intraductal therapy, and lithotripsy. In a 16-patient retrospective series after failed conventional extraction, cholangioscopy-guided laser lithotripsy achieved complete clearance in one session in every patient, but the sample was small and selected ^[149].

Cannulation failure and altered anatomy

Define difficult cannulation early rather than repeatedly traumatizing the papilla. Maintain wire access when possible, change the sphincterotome angle, use a pancreatic-guidewire-assisted approach when the pancreatic duct is entered, and consider a prophylactic pancreatic stent when repeated pancreatic access or precut therapy creates substantial risk. If standard access fails, expert options include needle-knife fistulotomy, needle-knife precut papillotomy, or transpancreatic sphincterotomy, selected according to papillary anatomy and operator experience. A rendezvous procedure, percutaneous or EUS-guided wire passage followed by endoscopic capture and ductal stenting, can restore access when retrograde cannulation is unsuccessful.

Do not persist with advanced cannulation when a safer drainage route is available. In a propensity-matched multicenter analysis of difficult cannulation in distal malignant obstruction, early EUS-guided biliary drainage had higher technical success than advanced ERCP techniques, 95.9% versus 82.0%, and fewer adverse events, 9.3% versus 18.2%; the retrospective design leaves residual selection bias possible ^[161]. Balloon-enteroscopy-assisted ERCP remains useful in Roux-en-Y or Billroth II anatomy. In a retrospective series of postoperative patients, double-balloon ERCP achieved single-session complete stone clearance in 94.8%, with adverse events in 2.8% ^[153]. EUS-guided treatment is another option for large stones in surgically altered anatomy, although matched retrospective data showed more sessions despite higher cumulative clearance than balloon-enteroscopy-assisted ERCP ^[152].

Strictures, hilar drainage, and device management

For a benign stricture, obtain tissue when the appearance is atypical or the diagnosis is uncertain, cross the narrowing with a guidewire, dilate conservatively, and use a removable stent strategy with a documented exchange or removal date. Multiple plastic stents remain a familiar approach when repeated remodeling and access are required. Fully covered self-expandable metal stents can shorten treatment in selected benign strictures but may migrate and can obstruct side branches; use them only when removal and ductal anatomy are secure. After living-donor liver transplantation, retrospective data found similar stricture-resolution and recurrence rates with multiple plastic, fully covered metal, and intraductal fully covered metal stents, while metal stents required fewer ERCP sessions and a shorter treatment duration ^[154].

For malignant obstruction, choose the stent according to resectability, expected survival, need for repeat access, side-branch anatomy, and whether tissue diagnosis remains pending. Plastic stents facilitate exchange and are useful when surgery or diagnosis is imminent; metal stents generally provide longer patency for unresectable obstruction. At the hilum, drainage is a three-dimensional planning problem rather than a simple choice of one versus two stents. Review cross-sectional imaging with the multidisciplinary team, target viable liver, avoid contaminating ducts that will not be drained, and use unilateral drainage when it adequately decompresses the relevant liver. Pursue bilateral or multi-segment drainage when unilateral drainage leaves a clinically important obstructed liver volume, provided each accessed sector can be drained reliably. ERCP technical success in malignant obstruction is high but repeat procedures remain common, so every stent placement requires a recorded device type, diameter, length, position, intended dwell time, and retrieval or exchange plan ^[151].

Use nasobiliary drainage when serial cholangiography is anticipated, the leak or infection requires direct monitoring, or temporary decompression is preferred before definitive stenting. Confirm free bile flow before ending the procedure. Remove or exchange plastic stents on schedule; an occluded or forgotten stent converts a therapeutic procedure into a source of recurrent cholangitis and repeat intervention. For occluded metal stents, endoscopic clearance, stent-in-stent placement, or selected intraductal radiofrequency ablation may be considered, but evidence for improved patency with radiofrequency ablation is inconsistent in randomized trials and remains heterogeneous ^[155].

Pearl: When diagnosis is the goal, use MRCP, CT, or EUS; when drainage or ductal therapy is the goal, ERCP is the intervention. Plan the endpoint before cannulation, and never place a stent without a retrieval, exchange, or definitive-patency strategy.

Therapeutic luminal endoscopy is definitive treatment when the target is intraluminal, technically reachable, and removable or bypassable without sacrificing oncologic clearance or organ safety. Choose endoscopy over surgery when complete eradication, durable hemostasis, or meaningful decompression is achievable; choose surgery when deep invasion, lymph-node clearance, diffuse disease, inaccessible anatomy, uncontrolled perforation, or failed endoscopic and radiologic rescue makes local therapy inadequate. Cross-sectional imaging and multidisciplinary review are preferable before attempting therapy for suspected transmural, extraluminal, ischemic, or metastatic disease. ^[168][176]

Endoscopic hemostasis

Treat active spurting or oozing bleeding and a nonbleeding visible vessel. For an adherent clot, ESGE 2026 suggests clot removal followed by treatment of the underlying vessel when safe; for high-risk ulcer bleeding, ESGE recommends contact or noncontact thermal therapy, mechanical therapy, or a sclerosant, either alone or with epinephrine injection. ^[163] Epinephrine produces temporary tamponade and vasoconstriction, so use it to slow brisk bleeding and expose the target, then add a definitive modality rather than using injection alone. Through-the-scope clips (TTSCs) are best for a discrete, capturable vessel or tear; rotate the clip, approach the vessel en face, capture sufficient tissue, and confirm closure. Fibrosis, tangential access, a large-caliber vessel, or a broad ulcer reduces clip reliability.

Use contact thermal therapy with a catheter or hemostatic forceps when the vessel can be compressed against the bowel wall; use noncontact argon plasma coagulation (APC) for superficial diffuse oozing, vascular ectasia, or a broad mucosal field. Apply the lowest effective energy and avoid prolonged focal activation because soft coagulation can cause deep tissue injury and delayed perforation. The AGA 2026 Clinical Practice Update describes electrosurgical advice as expert opinion without formal evidence grades. ^[166]

An over-the-scope clip (OTSC) grasps and compresses a larger volume of tissue and is useful for a fibrotic ulcer base, a large-caliber vessel, recurrent bleeding, or failure of standard therapy. ESGE 2026 suggests OTSC monotherapy as an alternative first-line treatment for actively bleeding high-risk peptic ulcers and recommends considering it for recurrent bleeding; persistent bleeding after all endoscopic modalities should proceed to transcatheter angiographic embolization (TAE), with surgery when TAE is unavailable or unsuccessful. ^[163] A 2025 meta-analysis of 14 studies, including five randomized trials and 7,211 patients, found lower 7- and 30-day rebleeding and 30-day mortality with first-line OTSC than with standard therapy, but the authors called for larger randomized trials. ^[173]

Hemostatic powders and gels are noncontact options when the bleeding point is diffuse, inaccessible, too large to capture, or obscured by brisk hemorrhage. They are particularly useful for malignant bleeding and as salvage or a bridge to embolization or surgery; do not regard powder as definitive treatment when a durable clip or thermal treatment is feasible. The 2026 international consensus was expert-led and explicitly identified the need for trials comparing powders. ^[165] For bleeding Mallory-Weiss tears and Dieulafoy lesions, the Canadian Association of Gastroenterology 2025 guideline conditionally favors band ligation or TTSCs, and mechanical therapy, contact thermocoagulation, or sclerosants, respectively, over epinephrine alone; for gastric antral vascular ectasia it conditionally favors band ligation over APC. ^[164] Endoscopic band ligation is also appropriate for esophageal varices and selected focal bleeding lesions when suction allows secure capture, but massive uncontrolled hemorrhage or failure of endoscopic control requires radiologic or surgical rescue. ^[179]

Resection of superficial neoplasia

Resect diminutive colorectal polyps with a cold snare that captures a rim of normal mucosa; avoid forceps except for truly diminutive 1-3-mm lesions when a snare cannot be deployed. Cold-snare polypectomy is preferred for most nonpedunculated lesions up to 9-10 mm and is reasonable for selected 10-20-mm sessile serrated lesions without dysplasia or invasion. A 2026 multicenter randomized noninferiority trial found complete histologic resection of intermediate-size sessile serrated lesions in 90.6% with cold snare, 88.5% with cold EMR, and 87.0% with hot EMR, although hot EMR achieved more en-bloc resections. ^[175] The evidence does not justify extrapolating cold snare to lesions with depression, ulceration, stiffness, disrupted surface pattern, or suspected submucosal invasion.

Use a hot snare for large pedunculated polyps, particularly when the stalk is thick or contains a prominent vessel; inject the stalk when needed, position the snare away from the muscularis, and transect with controlled electrosurgery. The AGA 2026 advice specifically favors hot rather than cold snare for large pedunculated polyps because thick stalk vessels may bleed when transected. ^[166] For a nonpedunculated lesion at least 20 mm without deep invasion, lift-assisted EMR is usually the efficient treatment. Remove the lesion en bloc when size and safety permit; otherwise use a planned piecemeal approach with complete peripheral clearance, retrieval of every fragment, and thermal treatment of the visible or resection margin when appropriate. AGA advises thermal ablation of margins after piecemeal colorectal EMR to reduce recurrence. ^[166]

Select ESD when en-bloc histology will determine curative treatment and EMR is unlikely to provide it: lesions with suspected superficial submucosal invasion, marked fibrosis or non-lifting, large lesions in which piecemeal resection would compromise staging, or lesions in anatomically difficult sites. ESD requires a circumferential incision, controlled submucosal dissection, hemostasis, and an intact specimen pinned for orientation. ASGE’s 2026 technology report emphasizes that ESD evidence often derives from case series and expert opinion because controlled trials remain limited. ^[167] Surgery is preferred for suspected deep submucosal invasion, lymphovascular invasion, poor differentiation, positive vertical margin, substantial tumor budding, or a lesion whose size, location, or fibrosis makes complete resection unsafe. ^[168][177]

Endoscopic full-thickness resection is reserved for selected small, localized, non-lifting or deep lesions when the defect can be closed and there is no evidence of advanced cancer. It is not a substitute for oncologic resection when lymph-node risk is material. ^[174][176] After any resection, the pathologist must receive an intact or clearly labelled piecemeal specimen with orientation, lesion size, depth, differentiation, lymphovascular invasion, and horizontal and vertical margins. An apparently complete endoscopic appearance cannot establish cure without histopathology. ^[168]

Luminal restoration, foreign bodies, and decompression

Dilate a benign stricture only after malignancy has been excluded when the morphology is atypical, ulcerated, or progressive. Use a wire-guided through-the-scope balloon with incremental diameter increases and stop for significant resistance, mucosal tearing beyond the intended effect, or severe pain. Refractory strictures may need repeated dilation, intralesional therapy, temporary stenting, or surgery; long, complex, ischemic, fistulous, or extrinsic strictures should prompt surgical or radiologic assessment rather than repeated blind dilation.

Place a self-expanding metal stent for symptomatic malignant esophageal, gastric outlet, or colorectal obstruction when rapid palliation is required or when it is an appropriate bridge to definitive surgery. Confirm the anatomy and length with cross-sectional imaging, cover the tumor with adequate margins, and document retrieval or exchange plans for removable devices. In selected benign obstruction, temporary fully covered stenting may be considered after dilation failure, but migration and tissue injury make surgery or radiologic approaches preferable when durable reconstruction is possible.

Remove esophageal or gastric food boluses, sharp objects, batteries, bezoars, and other foreign bodies urgently when they cause complete obstruction, airway risk, mucosal injury, or prolonged impaction. Use a retrieval net, snare, forceps, or protective hood according to shape and sharpness; extract batteries and sharp objects without delay and inspect the mucosa afterward. Decompress a markedly distended stomach or colon with suction, decompression tube placement, or therapeutic colonoscopy when the obstruction is functional or suitable for endoscopic decompression. Peritonitis, ischemia, perforation, closed-loop obstruction, or failed decompression requires immediate surgical or radiologic management.

Third-space and bariatric interventions

For achalasia, choose treatment after manometry, endoscopy, and assessment of esophageal morphology. Pneumatic dilation uses a graded 30-, 35-, and 40-mm balloon across the lower esophageal sphincter, inflated under fluoroscopic guidance at 7-15 psi for 15-60 seconds; it is most useful for patients who are poor candidates for general anesthesia, need temporizing treatment, or have recurrent symptoms after myotomy. ^[170] SAGES positions POEM or laparoscopic Heller myotomy as first-line therapy for type I and II achalasia and considers pneumatic dilation inferior to POEM for long-term symptom control; POEM is favored for type III disease because the myotomy can extend through the spastic segment. ^[170] POEM creates a submucosal tunnel and myotomy, then closes the mucosotomy with clips. Clinical success exceeds 90% at 7 years in the cited narrative review, but objectively detected reflux occurs in 40%-55%; give a proton-pump inhibitor after POEM and assess reflux objectively rather than relying on symptoms. ^[35]

Gastric peroral endoscopic myotomy (G-POEM) is used for selected refractory gastroparesis after objective confirmation of delayed emptying and exclusion of mechanical obstruction; it is particularly reasonable when symptoms persist despite dietary and medical therapy and the patient is treated at an experienced third-space center. POEM variants also include treatment of selected spastic esophageal disorders, Zenker diverticulum, and pyloric dysfunction. ^[35][172] Endoscopic suturing can close large mucosal defects, reinforce a leak or fistula repair, reduce a dilated gastrojejunal outlet, or revise selected bariatric anatomy. Use it when durable tissue apposition is required and the defect is reachable; severe fibrosis, inaccessible geometry, ongoing sepsis, or a large cavity may require drainage, radiologic intervention, or surgery.

After bariatric surgery, treat leaks and fistulas according to timing, cavity anatomy, tissue viability, and sepsis control. Options include covered bariatric stents, OTSCs for small well-defined defects, internal drainage for a mature cavity, septotomy for a chronic cavity with an accessible septum, and endoscopic suturing for selected defects. In a 2026 retrospective series of 61 patients, an algorithmic multimodal approach achieved clinical success in 90.2%, but complications occurred in 19.7% and larger self-limited cavities predicted failure. ^[171] Drain undrained collections percutaneously or surgically, control sepsis first, and refer for surgery when endoscopic access is impossible, the defect is large or ischemic, or endoscopic therapy fails.

Place a percutaneous endoscopic gastrostomy (PEG) for durable gastric feeding or decompression when oral intake is unsafe or inadequate and the gastrointestinal tract is functional. Place a percutaneous endoscopic jejunostomy (PEJ), or a jejunal extension through a PEG, when gastric feeding is unsuitable because of severe gastroparesis, recurrent aspiration, or gastric outlet obstruction. Confirm goals, expected duration, prognosis, and the feasibility of safe transillumination before placement; use radiologic or surgical placement when anatomy, ascites, interposed organs, altered surgical anatomy, or inability to obtain a safe puncture makes endoscopic access unreliable.

Pearl: Endoscopic therapy is curative only when the lesion is appropriately selected, completely treated, and validated by pathology or durable clinical decompression; when those conditions are absent, escalate early to expert resection, TAE, radiology, or surgery rather than repeating an unsafe procedure.

Tissue acquisition is successful only when the specimen answers the clinical question. Before sampling, define the lesion’s location, morphology, extent, and relationship to landmarks; record whether tissue is targeted, mapped, or removed en bloc. Send each anatomically distinct site in a separately labeled container, stating the site, lesion description, biopsy number, orientation, and specific diagnostic concern. ^[180]

For mucosal disease, use standard disposable forceps for most targeted biopsies. Jumbo forceps can provide a larger fragment, but depth and crush artifact remain limiting; they do not substitute for complete resection when invasion or margin assessment determines management. Place delicate mucosal fragments promptly in adequately filled 10% neutral-buffered formalin and avoid combining lesions or sites. If orientation matters, spread the fragment mucosal side up on filter paper or a biopsy pad before fixation; orientation is particularly valuable for small neoplasms and villous or dysplastic lesions because tangential sectioning can obscure architecture.

Use a snare when the lesion can be removed safely and the resulting specimen will provide more information than a superficial biopsy. A complete polypectomy, endoscopic mucosal resection , or endoscopic submucosal dissection specimen should be retrieved intact whenever possible, pinned or spread mucosal side up without stretching, and fixed before sectioning. The request should include lesion size, Paris morphology, resection method, whether removal was en bloc or piecemeal, and the orientation of any suspected deep or lateral margin. Histology then determines differentiation, depth of invasion, lymphovascular invasion, margin status, and whether endoscopic treatment is oncologically adequate. ^[9]

Brush cytology is complementary rather than interchangeable with tissue biopsy. It samples exfoliated surface cells and may help in a tight or difficult-to-biopsy stricture, but it cannot reliably assess invasion, stromal architecture, or the full immunophenotype. For an indeterminate biliary stricture, obtain ductal brushings and, when feasible, fluoroscopy-guided forceps biopsies or cholangioscopy-directed biopsies; place cytology material in the laboratory’s validated preservative and tissue fragments in formalin. Tell the pathologist whether the stricture is benign-appearing, indeterminate, or suspicious for cholangiocarcinoma, because a negative brushing does not resolve persistent radiologic or endoscopic concern.

EUS-FNA yields predominantly dispersed cells and is useful when cytology is sufficient, whereas EUS-FNB is preferred when preserved tissue architecture is needed, for example, suspected lymphoma , mesenchymal tumor, autoimmune disease, or a diagnosis requiring immunohistochemistry or molecular testing. ^[135] Use a dedicated pass for a cell block when lymphoma, metastatic disease, or a molecular assay is possible, and alert pathology before the procedure if fresh tissue is required for flow cytometry. Sample a subepithelial lesion only when the result will alter management; biopsy-related hemorrhage can be delayed and severe, as illustrated by a gastric plexiform fibromyxoma initially mistaken for a gastrointestinal stromal tumor . ^[185]

Immediate recognition determines whether an adverse event remains a transient procedural problem or becomes a time-critical emergency. Stop the procedure, call for help, remove the endoscope if it is worsening obstruction or aspiration, and assess airway, breathing, circulation, mental status, abdominal examination, bleeding, and procedural details. Give supplemental oxygen, attach continuous pulse oximetry and blood-pressure monitoring, obtain intravenous access, and use capnography when respiratory depression or deep sedation is present. ESGE/ESGENA recommends enhanced monitoring for high-risk patients and restricts reversal agents to sedation-related events that do not improve with nonpharmacological measures; patients should remain under trained observation until baseline observations return ^[20].

For apnea or severe hypoxemia, perform airway opening, jaw thrust, suction, bag-mask ventilation, and escalation to a supraglottic airway or tracheal intubation when ventilation remains inadequate or aspiration risk is substantial. Reduce or stop sedatives and opioids. Give flumazenil only for persistent benzodiazepine-related respiratory or central nervous system depression, because resedation and seizures can occur; give naloxone in small titrated doses when opioid-induced hypoventilation persists. Reverse respiratory depression, not merely slow recovery, and continue observation after apparent recovery. In obese patients, an oropharyngeal airway that preserves airway patency and permits capnography reduced hypoxia compared with a standard nasal cannula in a multicenter randomized trial ^[43].

Treat aspiration as a pulmonary emergency rather than as an automatic indication for antibiotics. Suction the oropharynx, provide oxygen and ventilatory support, and intubate when airway protection or gas exchange is inadequate. Obtain chest radiography or CT when hypoxemia, fever, respiratory distress, or a focal pulmonary finding persists. Start antibiotics for established aspiration pneumonia, sepsis, or severe immunocompromise, not for an isolated witnessed aspiration with rapid clinical recovery, and reassess the need for bronchoscopy with pulmonary or critical-care colleagues. Use piperacillin-tazobactam 4.5 g intravenously every 6 hours for 5-7 days when severe aspiration pneumonia requires broad hospital-acquired coverage; use ceftriaxone 2 g intravenously every 24 hours plus metronidazole 500 mg intravenously every 8 hours for 5-7 days when community-acquired aspiration pneumonia is clinically stable. Tailor therapy to cultures, renal function, local resistance, and the subsequent clinical course.

Severe abdominal pain, tachycardia, fever, peritoneal signs, subcutaneous emphysema, persistent vomiting, or unexpected hypotension after luminal endoscopy requires prompt evaluation for perforation , bleeding, ischemia, or obstruction. Stop insufflation, keep the patient nil by mouth, give intravenous crystalloid, analgesia, and broad-spectrum antibiotics, and obtain contrast-enhanced CT of the abdomen and pelvis; use water-soluble contrast orally or through the endoscope when a contained leak requires localization. Consult surgery early, even when endoscopic closure appears feasible. A small, recognized, contained defect without peritonitis may be closed with through-the-scope clips, an over-the-scope clip, endoscopic suturing, or a covered stent according to site and tissue quality; persistent leak, free perforation, peritonitis, sepsis, ischemia, distal obstruction, or failed closure requires operative management. EUS-guided hepaticogastrostomy deserves a low threshold for CT and surgical or interventional-radiology review when bile leak or peritonitis is suspected; in a retrospective series, early adverse events occurred in 15%, including peritonitis in 7% ^[26].

Bleeding may be immediate or delayed. Resuscitate with large-bore intravenous access, type and cross-match, serial hemoglobin and coagulation testing, and a restrictive transfusion strategy individualized to shock, ongoing hemorrhage, cardiovascular disease, and hemoglobin trajectory. Re-examine the patient and the procedure report to identify sphincterotomy, biopsy, dilation, polypectomy, EMR, ESD, PEG, or EUS puncture as the likely source. Repeat endoscopy is appropriate for ongoing or recurrent bleeding when the lesion is reachable. Use clips, contact thermal therapy, argon plasma coagulation for superficial diffuse oozing, or an over-the-scope clip for a large, fibrotic, recurrent, or refractory lesion; epinephrine injection is a temporizing measure and should be followed by definitive hemostasis. Failure or recurrence after endoscopic therapy warrants interventional radiology for transcatheter arterial embolization and surgical consultation if embolization is unavailable or unsuccessful. ESGE’s upper-GI-bleeding position statement defines successful hemostasis as absence of persistent or recurrent bleeding and requires referral pathways to surgery and interventional radiology ^[195].

Postpolypectomy bleeding is managed by hemodynamic assessment, bowel preparation or lavage as needed, and repeat colonoscopy for ongoing bleeding, hemoglobin decline, or transfusion requirement. Identify and treat the ulcer with clip closure, contact coagulation, or combination therapy; avoid epinephrine alone when a definitive modality is possible. After EMR or ESD, delayed bleeding is more likely with a large mucosal defect, exposed vessels, antithrombotic therapy, coagulopathy, or difficult hemostasis. Admit or observe patients with significant delayed hemorrhage, reverse anticoagulation only when the thrombotic risk permits, and involve the prescribing specialist. A multicenter retrospective study of esophageal ESD found post-ESD bleeding rates of 2.3% after guideline-based thienopyridine discontinuation and 3.2% with continuation, but these descriptive data do not establish that continuation is safe ^[199].

Post-EMR and post-ESD pain, fever, leukocytosis, or tachycardia can represent the expected inflammatory response, a deep mural injury, delayed perforation, or infection. Examine the abdomen repeatedly and obtain CT when symptoms are severe, progressive, or accompanied by systemic signs. A contained defect may be closed endoscopically; free perforation or deterioration requires surgery. Treat post-ESD coagulation syndrome or suspected transmural thermal injury with bowel rest, intravenous fluids, analgesia, and antibiotics when systemic inflammation or a contained injury is present, while excluding perforation before attributing symptoms to the syndrome.

Infection after endoscopy usually reflects contamination, incomplete drainage, or an infected target rather than routine diagnostic mucosal contact. Obtain blood cultures before antibiotics in a septic patient, culture drained fluid when available, and verify reprocessing records and traceability if an outbreak or device-related transmission is suspected. Current infection-prevention reviews emphasize high-level disinfection or sterilization of semicritical devices, objective process checks, traceability, and validated staff competency ^[198][208]. Give prophylactic antibiotics selectively for procedures with a substantial risk of bacteremia or infection, such as PEG placement, dilation of an infected or obstructed system, ERCP with incomplete drainage anticipated, and therapeutic EUS drainage; prophylaxis does not substitute for source control.

After ERCP, measure lipase or amylase when new epigastric pain, nausea, vomiting, or unexplained clinical deterioration occurs. Post-ERCP pancreatitis is diagnosed by compatible pain, enzyme elevation, and/or imaging according to accepted consensus criteria; manage it as acute pancreatitis with analgesia, early intravenous crystalloid guided by clinical response, prompt oral feeding when tolerated, and organ-supportive care. Escalate to monitored or intensive care for persistent hypotension, hypoxemia, renal failure, or other organ dysfunction. A retrospective 2026 cohort reported post-ERCP pancreatitis in 10.73% and associated it with difficult cannulation, pancreatic-duct opacification or guidewire access, prior pancreatitis, and sphincter-of-Oddi dysfunction; these predictors require external validation before use as a stand-alone clinical score ^[197].

Post-ERCP bleeding generally follows sphincterotomy and may present immediately or after discharge with melena, hematemesis, hematochezia, anemia, or shock. Resuscitate, correct clinically important coagulopathy, and perform repeat EGD/ERCP when bleeding persists. Endoscopic options include epinephrine followed by thermal or mechanical therapy, balloon tamponade of the papillotomy site, and covered metal-stent tamponade when standard therapy fails. Persistent bleeding proceeds to angiographic embolization and then surgery. Cholangitis requires cultures, intravenous antibiotics, fluid resuscitation, correction of reversible coagulopathy, and urgent biliary drainage when obstruction persists or the patient deteriorates. Use piperacillin-tazobactam 4.5 g intravenously every 6 hours for 4-7 days after source control, or ceftriaxone 2 g intravenously every 24 hours plus metronidazole 500 mg intravenously every 8 hours for 4-7 days when a narrower regimen is appropriate; extend or narrow treatment according to cultures and adequacy of drainage. If the stent is occluded or migrated, repeat ERCP for extraction, exchange, repositioning, or drainage of undrained segments; use percutaneous drainage or EUS-guided biliary drainage when ERCP fails. An international 2025 Delphi consensus recognized EUS-guided biliary drainage as an established salvage option after failed ERCP, but not yet as a universal first-line approach ^[196].

Fever and right-upper-quadrant pain after ERCP may also indicate cholecystitis , particularly when the cystic duct has been obstructed by contrast, sludge, or a stent. Obtain liver tests, blood cultures when febrile, and right-upper-quadrant ultrasound or CT. Give intravenous antibiotics and obtain urgent surgical consultation; percutaneous or EUS-guided gallbladder drainage is an alternative when surgery is unsafe. Stent migration, occlusion, fracture, or dislodgement requires imaging directed by symptoms and repeat ERCP or percutaneous retrieval when the device is accessible; do not leave a temporary plastic stent without a documented exchange or removal plan.

After PEG, inspect for bleeding, buried bumper, tube dislodgement, leakage, cellulitis, peritonitis, aspiration, and gastric or colonic injury. Stop feeds and obtain CT with water-soluble contrast through the tube when peritonitis, extravasation, or malposition is suspected. Treat superficial infection with local care and antibiotics directed by severity; deep infection, fascial involvement, abscess, or peritonitis requires surgery or interventional radiology. A dislodged tube should not be blindly replaced through an immature tract; confirm tract maturity and intragastric position before resuming feeds. In patients with ALS, lower respiratory reserve is associated with poorer outcomes after PEG, supporting respiratory assessment and anesthetic planning before placement ^[207].

Dilation can cause mucosal tearing, bleeding, perforation, or delayed chest or abdominal pain. Stop dilation when resistance is excessive or a deep tear appears, observe for instability or peritoneal signs, and obtain CT with water-soluble contrast when perforation is possible. Close a focal defect endoscopically when technically suitable; drain an associated collection and involve surgery for free leak, sepsis, or failed closure. EUS-guided drainage, biopsy, neurolysis, or transluminal intervention can produce bleeding, bile leak, perforation, peritonitis, pancreatitis, infection, or stent maldeployment. Doppler review before puncture, confirmation of a mature target and safe access route, and immediate CT or contrast evaluation for deterioration are essential; small-caliber bile ducts increase peritonitis risk after hepaticogastrostomy ^[26].

Capsule retention means that the capsule remains in the gastrointestinal tract for at least 2 weeks. Suspect it when the capsule is not seen in stool or on follow-up imaging, or when obstructive symptoms develop. Obtain abdominal radiography or CT, avoid further capsule passage attempts in symptomatic obstruction, and retrieve the capsule endoscopically or surgically according to location and cause. Before capsule use, investigate suspected obstruction or stricture with cross-sectional imaging and use a dissolvable patency capsule when uncertainty remains; a retained or failed patency capsule precludes diagnostic capsule ingestion. Retention risk is approximately 2% overall in the supplied small-bowel evidence, but is higher in stenotic or obstructing disease.

Ileus, distension, and nonspecific abdominal pain after endoscopy usually improve with observation, mobilization, correction of electrolyte disturbance, and minimizing opioids, but progressive distension, vomiting, obstipation, fever, or focal tenderness requires CT to exclude perforation, obstruction, ischemia, or bleeding. Decompress the stomach or bowel when vomiting or severe distension persists, and obtain surgical review for a mechanical obstruction or clinical deterioration.

An examination is incomplete when the intended landmark or therapeutic target is not reached, regardless of whether the limitation arose from poor preparation, intolerance, looping, fixed angulation, obstruction, stricture, equipment failure, or an adverse event. Document the indication, maximal extent reached, landmarks identified, preparation by segment, reason for termination, patient tolerance, sedation and rescue measures, lesions seen, biopsies or therapy performed, complications, images, and whether the unexamined territory remains clinically relevant. Do not imply completion from a limited examination. When safe, optimize cleansing, suction, water exchange, patient repositioning, abdominal pressure, a thinner or cap-assisted scope, or anesthesia support; do not force passage through a suspected malignant or high-grade stricture. Arrange repeat endoscopy with an intensified preparation or a different modality, or use CT/MR enterography, CT colonography, capsule endoscopy after patency assessment, interventional radiology, or surgery according to the unresolved clinical question. A complication-driven termination requires stabilization and an explicit rescue plan before any completion attempt.

Pearl: After any endoscopic adverse event, stabilize first, define the suspected injury with the least hazardous appropriate test, and escalate early when endoscopic rescue cannot provide reliable source control.

Quality assurance begins before the patient enters the procedure room. Confirm that the indication is appropriate, clinically specific, and recorded in the referral and procedure record; document the intended modality, urgency, relevant comorbidity, preparation status, consent, and planned intervention. Quality frameworks span appropriateness, preparation, consent, sedation, technical performance, documentation, adverse events, communication, and follow-up.^[58] A preprocedure checklist should reconcile the indication with the requested examination and identify when a different test, multidisciplinary review, or deferred procedure is safer. At the bedside, perform a documented timeout that confirms patient identity, procedure and extent, indication, allergies, anticoagulant or antiplatelet plan, sedation strategy, specimens, anticipated devices, and escalation plan. Record the names and roles of the endoscopist, assistant, nurse, and sedation professional; repeat the timeout after a change in operator, procedure, or intended therapeutic target.

Use a standardized report and retain the endoscopic images or video needed to prove examination extent and characterize abnormalities. Minimal Standard Terminology (MST) is a controlled vocabulary that makes findings searchable and comparable across operators and institutions; pair it with structured lesion documentation recording site, size, morphology, surface features, distance from landmarks, completeness of inspection, biopsy or resection method, retrieval, complications, and the responsible pathology specimen. Document whether resection was complete, en bloc or piecemeal, and whether residual tissue, uncertain margins, or an incomplete examination requires review. Give the patient and referring clinician a clear result, pathology plan, and follow-up responsibility; an electronic reminder or registry should close the loop rather than relying on an untracked recommendation. A Greek survey illustrates the practical gap: 71% of respondents did not record gastroscopy duration, only 62% reported routine photodocumentation, and complications were documented after colonoscopy in at least 95% of cases by 56%.^[58]

Quality indicators should be measured from consecutive eligible procedures, stratified by indication and endoscopist, and reviewed at regular multidisciplinary meetings. For colonoscopy , calculate cecal intubation rate from documented cecal landmarks, adenoma detection rate (ADR) as the proportion of screening or surveillance examinations containing at least one adenoma, and sessile serrated lesion detection rate separately because serrated lesions are not captured reliably by ADR alone. Record withdrawal time independently of insertion and therapy time; use a mean withdrawal time of at least 6 minutes as a practical benchmark for standard examinations, while recognizing that time is a process measure rather than proof of adequate mucosal inspection.^[58] AI-derived measurements may support audit, AI withdrawal time correlated strongly with manual measurement (r=0.91), but they remain an adjunct to clinical review.^[60]

Record bowel preparation by the Boston Bowel Preparation Scale (BBPS), with right, transverse, and left colon segment scores and the total score rather than a global label. A total BBPS of at least 6 with every segment at least 2 generally permits standard surveillance recommendations; a lower or segmentally inadequate examination should trigger documented rescue cleansing when feasible, explicit limitation of the examination, and a repeat or alternative plan.^[46] In a propensity-score-matched real-world cohort, water-exchange colonoscopy increased adequate preparation from 88.4% to 97.4% and sessile serrated lesion detection from 1.0% to 1.9%, but prolonged total procedure time from 15.6 to 22.6 minutes; these findings support measurement of both yield and workflow rather than adoption of a technique on detection alone.^[57]

For EGD , audit indication documentation, systematic examination of the required extent, landmark images, lesion characterization, biopsy or resection completeness, and appropriate follow-up assignment. A unit may use documented examination duration and landmark photodocumentation as process targets; the target should be defined in advance and monitored for every eligible examination rather than inferred from a normal report. For capsule endoscopy , record indication, ingestion and completion, small-bowel transit and visualization quality, retention or incomplete transit, lesion localization, image interpretation, report finalization, and documented referral for device-assisted enteroscopy or cross-sectional imaging when action is required. For EUS , audit indication, extent of anatomical assessment, lesion measurements and layer of origin, Doppler assessment before needle passage, needle type and passes, specimen adequacy, technical completion, adverse events, and whether the report leads to an appropriate multidisciplinary or pathology decision. For ERCP , measure indication appropriateness, selective biliary or pancreatic cannulation success, intended therapeutic success, stent traceability, repeat intervention, cholangitis, bleeding, perforation, and post-ERCP pancreatitis (PEP). Report PEP as a rate among ERCPs, using a prespecified clinical definition and severity grading; a retrospective cohort reported PEP in 19 of 332 procedures (5.7%), but its selective-indomethacin comparison was confounded by preferential treatment of higher-risk procedures.^[215]

A postprocedure conversation should answer four questions: what was examined, what was found, what tissue or imaging is still pending, and what happens next. Give the patient and referring clinician a written report that states the examination extent, limitations, preparation quality, lesions removed or biopsied, whether resection appeared complete, and the clinician responsible for reviewing pathology and closing the follow-up loop. Do not describe a colonoscopy as normal when the cecum was not reached or when inadequate preparation left a clinically relevant segment unseen. The report should reconcile endoscopic impressions with histology, cross-sectional imaging, EUS findings, and laboratory data; discordance is a reason to review the slides and images, obtain additional tissue or imaging, or discuss the case with the appropriate specialist rather than selecting the most reassuring result.

Before discharge, provide written instructions in plain language. State when the patient may eat, drive, work, resume antithrombotic therapy, and restart usual medicines; identify the expected degree and duration of pain, bloating, sore throat, or minor rectal bleeding; and give a telephone number that reaches the endoscopy service. Return urgently for increasing or severe abdominal or chest pain, persistent vomiting, fever or rigors, dyspnea, syncope, melena, hematochezia beyond the expected minor amount, jaundice, confusion, or inability to swallow. Delayed bleeding is possible after polypectomy, EMR, ESD, sphincterotomy, or ampullectomy, and delayed infection or obstruction can follow biliary intervention; the threshold for assessment should be lower in patients receiving anticoagulants, with major comorbidity, or after extensive resection. Any emergency contact should trigger review of the procedure report, medication changes, pathology status, laboratory results, and relevant CT, MRI, or ultrasound rather than an isolated symptom-based response.

A normal, complete, adequately prepared examination does not erase the indication. After a normal average-risk screening colonoscopy, the usual US Multi-Society Task Force (USMSTF) interval is 10 years; ESGE similarly returns patients with no relevant premalignant finding to the applicable screening program. Shorten that interval when the examination was performed for symptoms, when family history or hereditary risk changes the pretest probability, or when the patient has IBD or another disease-specific indication. A normal examination in Lynch syndrome is followed according to the pathogenic mismatch-repair gene and family risk rather than the average-risk schedule; in a retrospective cohort, MSH6 carriers developed adenoma or sessile serrated lesions more often during follow-up than MLH1 or MSH2 carriers. ^[101]

Do not assign routine surveillance from an inadequate examination. If the proximal colon was not adequately visualized, repeat colonoscopy as soon as practicable, generally within 1 year, with intensified preparation; an earlier repeat is appropriate when symptoms, iron-deficiency anemia, a positive FIT, or a suspected advanced lesion raises concern. If the examination is incomplete despite adequate preparation, document the reached landmark and unexamined segment, reason for termination, findings, and whether the limitation was anatomic, technical, tolerance-related, or caused by stenosis. Complete evaluation with repeat colonoscopy by an expert or CT colonography when appropriate; do not force a suspected malignant stricture. A positive CT colonography still requires colonoscopy for biopsy or removal when endoscopic therapy is feasible.

Pathology converts an endoscopic finding into a surveillance decision. The interval depends on histologic type, size, number, dysplasia, villous features, resection completeness, bowel preparation, family history, and comorbidity, not on the visual impression alone. Under USMSTF guidance, after a high-quality complete colonoscopy, use 7-10 years for 1-2 tubular adenomas smaller than 10 mm, 3-5 years for 3-4 such adenomas, 3 years for 5-10 adenomas, any adenoma at least 10 mm, villous or tubulovillous histology, or high-grade dysplasia, and 6 months after piecemeal resection of a lesion at least 20 mm. ESGE is less intensive for 1-4 small adenomas with low-grade dysplasia or small serrated lesions without dysplasia, generally returning such patients to screening, but recommends 3-year surveillance when the lesion burden or histology confers higher risk. Apply the shorter interval when the examination quality or completeness of excision is uncertain.

For sessile serrated lesions, confirm that the specimen was retrieved and that the pathologist can distinguish an SSL from a hyperplastic polyp. A completely excised SSL smaller than 10 mm without dysplasia generally returns to screening or a 5-10-year interval under USMSTF, whereas an SSL at least 10 mm or with dysplasia generally warrants 3-year surveillance; five or more SSLs smaller than 10 mm also warrant 3 years. Piecemeal removal of an SSL or adenoma at least 20 mm requires early scar inspection at approximately 6 months, followed by another examination about 12 months later if no residual tissue is found. Consider margin ablation documentation, photo comparison, and targeted biopsy of suspicious residual tissue; complete base ablation after EMR is intended to standardize the surveillance endpoint, but the cited 2026 report has no abstract available, so it does not establish a new interval. ^[201]

After colorectal cancer resection, surveillance is a coordinated oncologic program rather than repeated inspection of the anastomosis alone. Perform perioperative colon evaluation to identify synchronous lesions when feasible, then colonoscopy at 1 year after surgery, at 3 years if the 1-year examination is satisfactory, and every 5 years thereafter while the patient remains fit for treatment; shorten the interval if adenomas are found. Coordinate carcinoembryonic antigen testing and cross-sectional imaging with oncology and surgery according to stage and treatment plan. Refer to a colorectal surgeon when pathology shows noncurative endoscopic resection, such as deep submucosal invasion, lymphovascular invasion, poor differentiation, tumor budding, or involved margins, or when residual disease cannot be safely removed. Refer to oncology for systemic-treatment decisions and to a multidisciplinary tumor board when endoscopic, pathologic, and radiologic stage disagree, when nodal or metastatic disease is suspected, or when local excision competes with oncologic resection.

Barrett esophagus requires pathology-led surveillance after visible lesions have been resected or ablated. A nodular, ulcerated, depressed, or irregular area should undergo targeted sampling or endoscopic resection before surveillance is scheduled, because flat biopsies can under-stage focal neoplasia. For nondysplastic Barrett esophagus, use a 5-year interval for short-segment disease and 3 years for long-segment disease under current US guidance; confirm dysplasia with a second gastrointestinal pathologist and refer confirmed or suspected high-grade dysplasia or intramucosal cancer to an expert endoscopy center. After eradication therapy, document complete eradication of intestinal metaplasia and continue post-eradication surveillance because recurrence remains possible; the interval follows baseline dysplasia and the post-treatment protocol rather than the interval for untreated nondysplastic Barrett esophagus.

For gastric intestinal metaplasia, integrate the distribution of metaplasia, severity of atrophy, H. pylori status, family history, ethnicity or migration from a high-incidence region, smoking, and autoimmune gastritis. Test for and eradicate H. pylori, then confirm eradication; map the antrum/incisura and corpus separately when staging is needed. The AGA does not recommend routine surveillance for every patient with gastric intestinal metaplasia, but a 3-5-year interval can be considered after shared decision-making in patients with extensive or incomplete metaplasia, persistent H. pylori, family history, or other high-risk features; ESGE and British guidance generally favor approximately 3-year surveillance for extensive or advanced atrophy. Evidence syntheses of chronic atrophic gastritis address follow-up, H. pylori eradication, staging, endoscopic surveillance, medication, comorbidity, and special populations, supporting an individualized rather than purely histologic schedule. ^[218]

In IBD, surveillance begins after the duration and extent of colonic disease confer meaningful dysplasia risk and is performed with high-definition chromoendoscopy or enhanced imaging by an experienced operator. Remove a clearly visible, well-delineated, completely resectable lesion when safe; obtain expert pathology review. A surveillance interval of 1 year is appropriate for high-risk features such as prior dysplasia, stricture, primary sclerosing cholangitis, extensive inflammatory burden, or strong family history; 2-3 years is reasonable for intermediate risk, and up to 5 years may be used for sustained low-risk disease under disease-specific guidance. Shorten the interval when inflammation is active, preparation is poor, the lesion was incompletely removed, or histology and endoscopic appearance disagree. Crystal-violet delineation may assist selected UC-associated neoplasia, but it remains an adjunct supported by a four-case report: three cases had good pathological correspondence and one had a positive horizontal margin requiring additional ESD. ^[97]

After piecemeal EMR or ESD, the pathology report must state en-bloc versus piecemeal removal, lateral and deep margins when assessable, invasion depth, differentiation, lymphovascular invasion, and whether the specimen meets curative criteria. Inspect an EMR scar at 6 months, or earlier if residual tissue was suspected; after a negative examination, repeat at 12 months and then at 3 years, adapting the schedule to lesion site, pathology, resection method, and local protocol. Refer to an advanced endoscopist for residual or recurrent tissue that is nonlifting, fibrotic, poorly delineated, or not safely removable. Refer to surgery when pathology predicts lymph-node risk, margins remain involved after expert reassessment, or endoscopic retreatment cannot achieve oncologic clearance. In CDH1 pathogenic-variant carriers, endoscopic abnormalities can identify advanced signet-ring cell carcinoma but positive findings have modest positive predictive value; absence of abnormalities was reassuring in a retrospective consortium study, so surveillance versus prophylactic gastrectomy requires genetics, surgery, pathology, and multidisciplinary review rather than endoscopy alone. ^[32]

ERCP follow-up is part of the intervention, not an administrative afterthought. Record stent type, diameter, length, position, indication, insertion date, intended dwell time, and the exact exchange, removal, imaging, or laboratory trigger. Plastic biliary stents require a scheduled exchange or removal; do not leave the patient responsible for remembering an indefinite device. For benign strictures, leaks, or stones, schedule repeat ERCP or cross-sectional assessment according to duct healing and drainage; for malignant obstruction, align stent management with resectability, chemotherapy, radiation, and goals of care. A pancreatic duct stent placed for prophylaxis requires a documented confirmation of spontaneous passage or planned endoscopic removal; in a retrospective study, inappropriate retention of non-flanged stents occurred in 12.5% of patients with follow-up imaging, but only 2.2% required endoscopic removal. ^[217] Patients with fever, jaundice, worsening right-upper-quadrant pain, vomiting, or recurrent pancreatitis after ERCP need prompt laboratory testing and imaging, with gastroenterology and radiology or interventional radiology involvement when drainage is inadequate. Long-term transpapillary gallbladder stents also require active follow-up: a single-center retrospective series found late adverse events in 27% of high-surgical-risk patients, most commonly cholangitis or cholecystitis. ^[219]

When endoscopy is nondiagnostic, do not repeat the same examination reflexively. Review preparation, lesion targeting, specimen adequacy, pathology processing, and the concordance of endoscopic, radiologic, and laboratory findings. Persistent concern after negative biopsies warrants repeat targeted sampling, EUS-guided tissue acquisition, cross-sectional imaging, or referral to an expert center. Advanced disease, suspected nodal involvement, transmural extension, metastatic disease, an unresectable lesion, or a pathology result that changes staging should be discussed by gastroenterology, pathology, radiology, surgery, and oncology in a multidisciplinary tumor board. EUS-guided treatment of metastatic pancreatic lesions illustrates the required discipline: a small retrospective series treated only histologically confirmed metastases in a multidisciplinary setting, and the authors emphasized the need for larger prospective multicenter studies. ^[139]