Subscribe
Share
Share
Embed
These patients share one principle: the mucosa was injured by something that touched it, whether an organism the host could not keep down, a tablet that lodged where the lumen narrows, or a swallowed corrosive. Depth reflects contact time, site reflects local anatomy, and host status reflects who is vulnerable. That framework sorts the differential before the scope.
The case. An HIV patient with a CD4 count of 40 has odynophagia and one large, deep, serpiginous distal ulcer on endoscopy. Where do you biopsy, and what organism are you looking for?
Topics covered
Key decisions
For the full chapter with MCQs, tables, and primary-guideline references, visit www.boardpearls.com.
Questions or feedback: hello@boardpearls.com.
Board Pearls is a gastroenterology board review built around clinical reasoning, not recall. Each episode takes one high-yield topic and works it the way you would on rounds: a case to anchor it, the framework that sorts the differential, and the specific decisions the exam rewards.
The gastroenterology series covers the full blueprint across nine modules: esophagus, stomach and duodenum, small bowel, colon, pelvic floor, liver, pancreas and biliary, endoscopy, and the cross-cutting topics. Episodes are grouped by chapter and built from the primary guidelines and pivotal trials the boards draw from (ACG, AGA, AASLD, ASGE), not from textbook summaries.
Use it as an audio companion to the written curriculum, MCQs, and AI tutor at boardpearls.com. Questions or feedback: hello@boardpearls.com.
Welcome to Board Pearls. This is episode two of two of the Eosinophilic and Infectious Esophagitis chapter, in the Esophageal Disorders module. In this episode we cover the infectious and direct-injury esophagitides: Candida as the most common infectious cause and the gateway to immunocompromise workup, HSV with shallow volcano-edge ulcers and epithelial inclusions at the edge versus CMV with deep distal ulcers and endothelial inclusions at the base, pill esophagitis as contact injury at sites of esophageal narrowing, and caustic injury graded by the Zargar endoscopic scale.
The patients in this episode are linked by a single principle. The esophageal mucosa is injured by something that touched it. The something is an organism that the host's defenses failed to keep down, a tablet that lodged where the lumen narrows, or a swallowed corrosive. The depth of the injury reflects the contact time. The site of the injury reflects local anatomy. The host context tells you which agent is most likely, and the biopsy site tells you where the agent lives in the wall. Once those three axes are in mind, the differential sorts itself.
Start with Candida, because Candida is the most common infectious esophagitis and is the diagnosis the boards reach for first when a patient with an immune-compromise risk factor presents with odynophagia. The mechanism is overgrowth of yeast that already lives on the mucosa. Candida albicans is part of the normal oral and esophageal flora in many adults. It is kept down by salivary antimicrobial peptides, by gastric acid, by propulsive motility, and by intact cell-mediated immunity. When any of those defenses fail, the yeast proliferates and invades the squamous epithelium.
The host pattern is the diagnostic spine. HIV with a CD4 count under two hundred cells per microliter is the canonical setting, and esophageal candidiasis can be the AIDS-defining illness in a previously healthy adult. That is why an HIV test belongs in the workup of any unexpectedly severe Candida esophagitis without an obvious explanation. Solid-organ and stem-cell transplant recipients on calcineurin inhibitors are at risk. Chronic systemic steroid users are at risk, and so are inhaled-steroid users. The asthmatic on a fluticasone inhaler and a PPI is the canonical primary-care vignette, because the inhaled steroid suppresses local mucosal immunity and the PPI removes the acid barrier. Poorly controlled diabetes, hematologic and solid malignancy, and esophageal stasis from achalasia or scleroderma or a peptic stricture all qualify. Long-term broad-spectrum antibiotics shift the flora and remove bacterial competition. The PPI itself is a contributor, because acid is one of the constitutive antifungal barriers.
The presentation is odynophagia, dysphagia, retrosternal discomfort, and sometimes weight loss from reduced intake. Oral thrush is present in many but not all patients, and its absence does not exclude esophageal involvement, especially in inhaled-steroid users who rinse their mouths after dosing. Endoscopy shows discrete white or yellow-white plaques adherent to the mucosa, typically in the mid and distal esophagus and often diffuse. The plaques do not wash off with water irrigation, which separates them from food residue. They can be pushed but not easily removed, which separates them from leukoplakia. Brushings or biopsies of the plaques show pseudohyphae and yeast forms on Gram, periodic acid-Schiff, or Grocott methenamine silver staining. Pseudohyphae are the histologic fingerprint of Candida invasion. True hyphae would prompt reconsideration toward Aspergillus or another mold.
Treatment is fluconazole at two hundred to four hundred milligrams orally daily for fourteen to twenty-one days. The dose is higher than for oral thrush because the esophageal mucosa needs higher tissue concentrations to clear invasion. The intravenous route at the same dose range is used when the patient cannot swallow because of severe odynophagia or has impaired absorption. Echinocandins, which include caspofungin, micafungin, and anidulafungin, are reserved for fluconazole-refractory disease, for severe disease in the critically ill, and for documented or suspected azole-resistant species. Candida glabrata in particular is increasingly fluconazole-resistant and is the species the boards expect a candidate to name in the transplant patient who has already failed an azole. Voriconazole and posaconazole are the alternatives in resistant cases. Amphotericin B is rarely needed in modern practice. Maintenance prophylaxis is not routine after a single episode. In HIV with persistently low CD4 counts the durable prevention is immune reconstitution through antiretroviral therapy, not chronic fluconazole, which selects for resistance.
There is a small set of conditions that produce a white esophagus and that the candidate should hold next to Candida at the moment of endoscopy. Leukoplakia is non-removable, rough squamous keratinization associated with chronic injury and tobacco. Esophagitis dissecans superficialis presents as sloughing sheets of squamous epithelium that lift off in strips, and it is associated with bisphosphonates, celiac disease, autoimmune blistering disease, and some psychotropic medications. Pseudomembranous CMV esophagitis can produce extensive adherent exudate that overlaps endoscopically with severe Candida, and the discriminator is biopsy showing stromal endothelial inclusions rather than pseudohyphae. Parakeratosis presents as white mucosa associated with head and neck cancer or chronic injury. The diagnostic step in any white-esophagus presentation is biopsy or brushing. An empiric fluconazole trial without endoscopy is reasonable in a low-risk patient with classic plaques and oral thrush. A patient with severe odynophagia, immunocompromise, or atypical findings should go directly to EGD with biopsy. The point is to rule out a viral or pill-induced process before the empiric antifungal is started.
That brings us to the viral esophagitides, and the sorting principle here is one of the cleanest mechanistic distinctions in esophageal pathology. HSV is an epithelial infection. The virus replicates in the squamous epithelial cells. CMV is a stromal infection. The virus replicates in endothelial cells and stromal fibroblasts of the submucosa. That single fact dictates everything downstream. It dictates the ulcer depth, because epithelial infection produces shallow ulcers and stromal infection produces deep ulcers. It dictates the biopsy site, because viable infected cells live at the ulcer edge for HSV and at the ulcer base for CMV. It dictates the histologic fingerprint, because the inclusions appear in different cell types. And it sets up the most common careful-reasoning failure on this material, which is biopsying the CMV ulcer edge instead of the base and getting back granulation tissue without diagnostic inclusions.
HSV esophagitis is recognized by multiple shallow, well-circumscribed ulcers with raised, rolled edges, often described as volcano-edge or volcanic-eruption ulcers. The lesions are scattered through the mid and distal esophagus. Each ulcer typically begins as a small vesicle that ruptures and leaves a shallow defect with a heaped-up margin and a clean fibrinous base. The histologic fingerprint, present in the squamous epithelial cells at the edge of the ulcer, is the trio of multinucleation, margination of chromatin against the nuclear membrane, and molding of nuclei against one another. Intranuclear Cowdry type A inclusions appear as eosinophilic globules surrounded by a clear halo. Multinucleated giant cells with these features are pathognomonic. Immunohistochemistry for HSV-1 and HSV-2 confirms when histology is suggestive but not definitive. Viral culture or PCR of biopsy material is supportive.
The biopsy has to come from the heaped-up border. The ulcer base is granulation tissue without viable infected epithelial cells, and a base biopsy returns nondiagnostic tissue. The candidate who biopsies the edge surfaces the diagnosis. The candidate who biopsies the base does not.
The host pattern for HSV is broader than for CMV. Solid-organ and stem-cell transplant recipients, chemotherapy patients, and other immunocompromised hosts are the classic populations. HSV also occurs in immunocompetent young adults, often in the setting of recent oral or labial herpes, with a self-limited course in that population. Treatment in the immunocompromised host is acyclovir at four hundred milligrams orally five times daily for fourteen to twenty-one days. Intravenous acyclovir at five to ten milligrams per kilogram every eight hours is used for severe disease or when oral intake is impossible. In the rare immunocompetent case, acyclovir at two hundred milligrams orally five times daily for seven to ten days is sufficient. Four hundred milligrams three times daily for the same duration is the alternative dosing. Many cases resolve without antivirals if the patient can be supported through the odynophagia. Foscarnet is reserved for acyclovir-resistant HSV, which is increasingly recognized in heavily immunosuppressed patients with prior acyclovir exposure.
CMV esophagitis is recognized by one or more large, deep ulcers in the distal esophagus, sometimes solitary and sometimes multiple, often serpiginous or geographic in shape. In severe cases the ulcers coalesce into broad areas of denuded mucosa. The ulcers are deeper and more extensive than HSV ulcers, and they tend to sit in the distal third. The histologic fingerprint is the cytomegalic cell. This is a markedly enlarged endothelial cell or stromal fibroblast in the granulation tissue at the base of the ulcer. The cell contains a prominent intranuclear basophilic inclusion surrounded by a clear halo, the classic owl-eye, and an additional ground-glass intracytoplasmic inclusion that is more variable. The inclusions sit in stromal cells, not in the surface epithelium. That is the diagnostic counterpoint to HSV. It is the reason the CMV biopsy is taken from the ulcer base.
Immunohistochemistry for CMV antigens confirms the diagnosis. CMV PCR of blood supports a diagnosis of systemic CMV disease but does not by itself establish esophageal involvement, because a viremic patient can have a different cause for the esophageal lesion. A negative blood PCR does not exclude tissue-invasive disease either. The diagnosis lives in the biopsy.
The host context points to CMV in HIV with a CD4 count below one hundred cells per microliter. It also points to solid-organ transplantation particularly in the first six months, when reactivation is most common. Stem-cell transplantation and chemotherapy-induced neutropenia round out the high-risk hosts. Treatment is intravenous ganciclovir at five milligrams per kilogram every twelve hours as induction for fourteen to twenty-one days. Transition to oral valganciclovir at nine hundred milligrams daily completes the course. Maintenance suppression continues in patients who require it because of ongoing immunosuppression. Foscarnet is the alternative for ganciclovir-resistant disease and for patients in whom ganciclovir-induced myelosuppression is dose-limiting. Oral valganciclovir alone may be reasonable for milder cases in selected outpatient transplant recipients with adequate absorption, but the standard induction in moderate to severe disease is intravenous ganciclovir. Any patient with documented invasive CMV should have an ophthalmology evaluation, because CMV retinitis is a common concomitant infection and is sight-threatening if missed.
The clean way to hold all of this is the epithelial-versus-stromal axis. HSV ulcers are shallow with volcanic raised edges, the biopsy comes from the edge, the histology shows multinucleated giant cells with Cowdry A inclusions in epithelial cells, and the treatment is acyclovir. CMV ulcers are large and deep with broader geographic morphology. The biopsy comes from the base. The histology shows owl-eye intranuclear inclusions plus cytoplasmic inclusions in stromal endothelial cells and fibroblasts. Treatment is intravenous ganciclovir induction followed by oral valganciclovir maintenance. A patient with both viruses on biopsy, which does occur in heavily immunosuppressed hosts, gets both regimens. A patient with negative biopsies but high clinical suspicion in an immunocompromised host can be treated empirically while pathology and immunohistochemistry are pending, particularly if odynophagia is severe.
Now to the direct-injury esophagitides. The mechanistic principle that links them is the same principle that organized the infectious workup: contact time equals depth. The agents differ. A slowly dissolving tablet is the chemical in pill esophagitis, and a corrosive liquid is the chemical in caustic injury. The contact times differ. Minutes for a lodged pill, seconds to minutes for a swallowed caustic. But in both, the wound is determined by how long the agent sat on the mucosa and what kind of chemistry it brought.
Pill esophagitis is direct mucosal injury at the point where a tablet lodged against the esophageal wall and dissolved in place. The mechanism is local toxicity from the dissolved drug. Low pH does it for doxycycline, tetracycline, ascorbic acid, and ferrous sulfate. High osmolality does it for potassium chloride extended-release. Direct cytotoxicity does it for NSAIDs, aspirin, bisphosphonates, dabigatran, quinidine, and clindamycin. The injury runs from a coagulative or apoptotic mucosal surface lesion into the lamina propria depending on contact time. The classic offenders the boards expect a candidate to recognize fall into a short list. Doxycycline and tetracycline lead the list because they are widely prescribed and acidic when dissolved. Potassium chloride is next, especially in extended-release formulations. The oral bisphosphonates are alendronate, risedronate, and ibandronate. NSAIDs and aspirin are common contributors. Ferrous sulfate, ascorbic acid, clindamycin, quinine and quinidine, and the direct oral anticoagulant dabigatran round out the list.
The risk factors are pharmacologic and behavioral. Taking a pill with little water leaves the tablet poorly hydrated and increases the likelihood that it adheres to the mucosa as it dissolves. Taking a pill while supine, or shortly before lying down, allows gravity to fail as a propellant. Underlying motility disorders, including achalasia, scleroderma, and ineffective esophageal motility, prolong contact time at the abnormal segment. Pre-existing strictures and rings do the same. Older adults are overrepresented because of polypharmacy, reduced saliva, and frequent recumbency.
The presentation is sudden onset of retrosternal odynophagia, often with associated dysphagia and chest pain, hours to days after an inciting pill. The classic stem is a patient who took the medication at bedtime with a small sip of water. The pain is sharp and worse with swallowing. The history is the diagnostic key. The candidate who asks specifically about pill timing, water volume, and posture will surface the etiology before endoscopy. Endoscopy when performed shows discrete, often shallow ulcers with normal intervening mucosa. The ulcers sit most commonly at the level of the aortic arch, where the left mainstem bronchus indents the esophagus and a tablet can lodge. They also occur at the lower esophageal sphincter, where peristalsis slows. The kissing-ulcer pattern, in which paired ulcers face each other on opposing walls because the tablet sat between them, is the classic morphology. Histology shows ulceration with a mixed inflammatory infiltrate including neutrophils and reactive epithelial change, without the eosinophilic pattern of EoE or the viral inclusions of HSV and CMV.
Treatment is removal of the offending agent, sucralfate slurry to coat the injured mucosa and accelerate healing, and brief PPI therapy when concurrent reflux is contributing or the ulcer is severe. Topical viscous lidocaine and systemic analgesia help in the acute phase. The crucial intervention is prevention through swallowing-technique counseling. Take pills upright with at least eight ounces of water. Do not lie down for at least thirty minutes after the dose. Consider alternate formulations or routes for high-risk medications in patients with motility disorders or strictures. Symptoms typically resolve within one to two weeks. Recurrent pill esophagitis in the same patient warrants reassessment for an underlying motility disorder or stricture that prolongs contact time.
Caustic injury is the result of a swallowed corrosive, and the depth of the injury is dictated by the chemistry of the agent. Strong alkali produces liquefactive necrosis. The agents are sodium hydroxide and potassium hydroxide, found in liquid drain cleaners, oven cleaners, and some industrial cleaning products. Lye is the colloquial term for sodium hydroxide. The mechanism is saponification of membrane lipids and dissolution of structural proteins. The alkali penetrates rapidly and deeply through the mucosa, through the muscularis, and toward the mediastinum until tissue fluids buffer it. Alkali ingestion preferentially damages the esophagus. The agent passes through the oral cavity and pharynx quickly and concentrates in the esophageal lumen. Gastric acid partially neutralizes the alkali once it reaches the stomach, which limits gastric injury.
Strong acid produces coagulative necrosis. The agents are hydrochloric, sulfuric, and phosphoric acids, found in toilet bowl cleaners, swimming pool chemicals, and some battery products. The mechanism is denaturation of proteins, which forms a coagulated eschar at the surface that limits further penetration. Acid transits the esophagus faster, which biases the injury toward the stomach and the pylorus, where pooling and tighter outflow concentrate the agent. The generalization to retain is that alkali damages the esophagus and acid damages the stomach. That generalization carries most stems, even though severe ingestion of either can produce injury at any level.
Two scenarios sit outside that generalization. Liquid household bleach at five percent sodium hypochlorite is a frequent ingestion in pediatric and self-harm contexts. It rarely causes severe esophageal injury. The concentration is too low and the compound is more irritant than caustic. Disc battery ingestion in children is a separate emergency. Local electrolytic generation of hydroxide at the negative pole produces rapid alkaline injury within hours, and the management is urgent endoscopic removal, which is detailed in the GI emergencies chapter.
Initial management of caustic ingestion is airway, breathing, circulation, and the airway is the immediate priority. Oropharyngeal and glottic edema can progress rapidly. Intubation should be performed early in any patient with stridor, drooling, voice change, or extensive oropharyngeal burns. The patient is kept NPO. Intravenous access, fluid resuscitation, and analgesia are established. Three things are explicitly contraindicated, and the candidate who knows why does not have to memorize the list. Emetics are contraindicated because vomiting re-exposes the esophagus to the agent and can drive perforation in already weakened tissue. Neutralizing agents are contraindicated because the neutralization reaction is exothermic, and the heat compounds the chemical injury. Nasogastric tube placement is contraindicated initially because it can induce retching and vomiting and can perforate already injured mucosa. Placement under endoscopic guidance during the staging endoscopy is acceptable in selected patients when nutritional access is needed. Activated charcoal is not effective and is not given. Antibiotics are not routine unless perforation or aspiration is suspected.
Endoscopy is performed within twenty-four to forty-eight hours of ingestion to establish the depth and extent of injury, and the findings are graded by the Zargar classification. Zargar grade zero is normal mucosa. Grade one is mucosal edema and hyperemia. Grade two-A is superficial ulceration with bleeding, blisters, or whitish exudate. Grade two-B is deep focal or circumferential ulceration. Grade three-A is focal necrosis. Grade three-B is extensive necrosis. Grade four is perforation. The grade dictates disposition and prognosis. Grades zero and one are managed with a liquid diet advanced as tolerated, observation, and outpatient follow-up. Grade two-A is similar with a longer observation period. Grade two-B and grade three require ICU admission, NPO status, parenteral nutrition or nasojejunal feeding placed under endoscopic guidance to bypass the injured esophagus, and surgical consultation. Grade three-B in particular is a high-mortality lesion with substantial transmural necrosis, and esophagectomy is sometimes required for non-viable tissue or established perforation. Grade four is established perforation and is a surgical emergency. Endoscopy is contraindicated in patients who are hemodynamically unstable, who have evidence of perforation, who have severe respiratory distress, or who have severe oropharyngeal or glottic necrosis that makes safe instrumentation impossible. In those patients computed tomography is the imaging modality of choice and surgery is consulted directly.
Pharmacologic management of caustic injury has settled into a narrow set of evidence-based interventions. PPI therapy reduces gastric acid contribution to ongoing mucosal injury and facilitates healing. Sucralfate is sometimes used as an adjunct. Systemic corticosteroids have been studied repeatedly for stricture prevention in grade two-B injury. The evidence does not support routine use. Some centers continue to use steroids selectively for grade two-B injury based on individual benefit-risk. Steroids are not used in grade three because they impair healing and increase infection risk. Antibiotics are reserved for documented or strongly suspected perforation and aspiration. Mitomycin C applied topically during early dilation has been used to reduce stricture recurrence in some pediatric series, but it is not standard adult care.
The long-term sequelae of caustic injury are stricture and squamous cell carcinoma. Esophageal stricture develops in roughly a third of patients with grade two-B injury and in most patients with grade three injury, typically within two to eight weeks as the mucosa heals with fibrosis. Stricture management is graduated endoscopic dilation, sometimes serial over months to years, with steroid injection or stenting in refractory cases. Esophagectomy with reconstruction is reserved for non-dilatable strictures and for patients with intolerable nutritional consequences. Esophageal squamous cell carcinoma develops with long latency on the order of twenty to forty years after the original injury. The relative risk is on the order of a thousand-fold compared with the general population in some series. That is why surveillance endoscopy beginning roughly fifteen to twenty years after injury, and continuing thereafter, is recommended in patients with significant initial injury. The mechanism is chronic inflammation, fibrotic scarring, and dysplasia in the regenerating squamous mucosa. The clinically load-bearing point is that caustic ingestion is not just an acute emergency. It is a lifelong premalignant condition for which the patient needs counseling, surveillance, and indefinite engagement with GI care.
Pull the episode together on the axis it began on. The injured mucosa tells you the agent if you read three things together. The host context narrows the infectious differential before endoscopy. Candida fits the steroid-and-PPI patient and the HIV patient with a CD4 under two hundred. Viral disease fits the deeper immunocompromise of transplant, neutropenia, and HIV with a CD4 under one hundred. The ulcer morphology tells you the organism. Shallow volcano-edge ulcers are HSV. Deep distal serpiginous ulcers are CMV. Discrete shallow ulcers at the aortic arch with normal intervening mucosa are a pill. White plaques that do not wash off are Candida. The biopsy site has to match the cellular tropism. Edge for HSV epithelial cells, base for CMV stromal cells, plaque for Candida pseudohyphae. And in the chemical injuries, contact time and chemistry set the depth. Alkali liquefies and goes deep into the esophageal wall. Acid coagulates and biases toward the stomach. The Zargar grade tells you whether the patient goes home, goes to a ward, goes to the ICU, or goes to the operating room.
The next chapter turns to peptic ulcer disease and to Helicobacter pylori, where the question is no longer which agent injured the mucosa but how to clear an organism that has become harder to kill. Bismuth quadruple therapy and PCAB-based regimens have replaced clarithromycin triple therapy as first-line in most populations. Salvage regimens are built around avoiding drugs the patient has already seen. NSAID-associated ulcer risk is multiplicative rather than additive, and it is prevented by combining a PPI with eradication of H. pylori when both factors are present.