Subscribe
Share
Share
Embed
Painful swallowing and a failed proton pump inhibitor trial share one idea: the symptom tells you what kind of problem you have. Odynophagia means injury to the lining, so the episode asks about pills, immune status, and prior reflux, caustic ingestion, or radiation, then separates Candida, herpes, and cytomegalovirus, decides between empiric treatment and endoscopy, and aims biopsies at the ulcer edge or base. Acute radiation mucositis follows. The second half treats a PPI failure with a normal workup as a result to interpret, working through rumination, globus, and functional chest pain, and ending with the rule to rethink the diagnosis rather than raise the dose.
Topics covered
Key decisions
This is an AI-generated podcast, and some pronunciations may be imperfect. Thank you for your understanding, and we hope you enjoyed this content.
For the full chapter with MCQs, tables, and primary-guideline references, visit www.boardpearls.com.
Questions or feedback: hello@boardpearls.com.
The most complete gastroenterology board-review podcast on the web, and it is entirely free. This is the GI series from Board Pearls: board review built around clinical reasoning, not recall. Every 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. No trivia, no textbook dictation, just the thinking that earns the point.
Over 100 episodes span 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), grouped by chapter and built from the guidelines and pivotal trials the boards are written from: ACG, AGA, AASLD, and ASGE.
Pair it with the written curriculum, question bank, and AI tutor at boardpearls.com. Questions or feedback: hello@boardpearls.com.
Welcome to Board Pearls. This is episode two of three of the Esophageal Symptoms and Diagnostic Workup chapter, in the Esophageal Disorders module. This episode covers two problems, through five patients. The first problem is swallowing that hurts. A forty-one-year-old kidney transplant recipient with cytomegalovirus esophagitis teaches what pain means about the lining, how three history clues narrow the cause, when to treat empirically and when to scope, and where on an ulcer to biopsy. A fifty-eight-year-old woman in the middle of chemoradiation teaches acute radiation mucositis. The second problem is the patient who seems to have reflux, doesn't improve on a proton pump inhibitor, PPI, and has a normal workup. A twenty-four-year-old woman with rumination syndrome, a fifty-two-year-old man with globus, and a forty-five-year-old woman with functional chest pain teach how to recognize that the story isn't reflux, how each diagnosis is confirmed, and why a higher PPI dose is the wrong answer.
Here's the first patient. A forty-one-year-old man received a kidney transplant eight months ago and takes tacrolimus, mycophenolate, and prednisone. For two weeks, swallowing has hurt: a sharp pain behind his breastbone with every bite, and now with every sip. He's eating very little and has lost six pounds. He's never had heartburn, he hasn't started any new pills, and he's never swallowed anything caustic. He's had no chest or neck radiation.
His symptom is odynophagia, and its meaning sets the whole workup. Dysphagia is food sticking. Odynophagia is pain on swallowing. A narrowing or a weak muscle makes food stick, but it doesn't make swallowing hurt. Pain means the lining of the esophagus is injured: an ulcer, an infection, or a chemical burn. So odynophagia is treated as mucosal injury until proven otherwise, and the real question is what injured the lining.
Three history clues identify the injury in most patients. Was a pill swallowed with too little water or just before lying down? Is the patient immunocompromised: human immunodeficiency virus, HIV, with its CD4, cluster of differentiation four, count, a transplant, chemotherapy, or chronic inhaled or systemic steroids? And is there prior reflux, a caustic ingestion, or radiation? Each answer points to a different injury, and together they usually narrow the cause to one or two diagnoses before the endoscope goes in, so the endoscopy can be aimed at the right biopsy site.
Pill injury is asked about first, because it's often diagnosed from the history alone. Some drugs are acidic, hyperosmolar, or directly toxic when they dissolve against the lining: doxycycline and tetracycline, potassium chloride, bisphosphonates, nonsteroidal anti-inflammatory drugs, NSAIDs, iron, and dabigatran, among others. If a tablet lodges in the esophagus, it injures the lining where it sits. The usual site is the level of the aortic arch, where the left mainstem bronchus indents the esophagus and slows the pill down. The story is sudden odynophagia, hours to days after taking that specific pill with a sip of water or at bedtime. When endoscopy is done, it shows discrete shallow ulcers, often a pair of kissing ulcers on opposite walls, with normal mucosa between them. Treatment is stopping the drug, sucralfate to coat the ulcer, and teaching the patient to take pills upright with a full glass of water.
He took no new pill. The second clue decides his case: he's on three immunosuppressants, and in an immunocompromised patient odynophagia is usually infection. His negative reflux, caustic, and radiation history leaves nothing competing with that explanation.
Three infections cause esophagitis in an immunocompromised host, and the distinction matters because each looks different at endoscopy, lives in a different layer of the tissue, and needs a different drug. Candida is the most common. It appears with inhaled or systemic steroids, PPI use, HIV with a CD4 count under two hundred, transplantation, antibiotics, and esophageal stasis, where food sits. At endoscopy it's adherent white plaques, and brushings show pseudohyphae. Fluconazole is first line.
The two viruses differ in the cells they infect. Herpes simplex makes shallow ulcers with raised, volcano-like edges. The virus infects the squamous epithelial cells of the lining, so its inclusions sit at the ulcer's edge, and the pathologist sees multinucleated giant cells. Acyclovir treats it. Cytomegalovirus makes large, deep ulcers, often solitary, typically in the distal esophagus, in transplant recipients and in HIV with a CD4 count under one hundred. It infects the stromal and endothelial cells beneath the lining, so its owl's-eye inclusions sit in the base of the ulcer. It's treated with intravenous ganciclovir, followed by oral valganciclovir.
A transplant recipient is at risk for all three. Candida is the most common overall, cytomegalovirus is the classic infection after transplant, and herpes is possible too. Each needs a different drug, so the most likely one can't simply be treated blind.
Because the three need different drugs, the next decision is whether a patient can be treated empirically or needs a scope now. An empiric course of fluconazole is reasonable in a mildly symptomatic patient with a clear Candida risk factor, such as a PPI user on an inhaled steroid who has thrush; a quick response to fluconazole points to Candida without an endoscopy.
Four situations go straight to endoscopy: severe odynophagia, an immunocompromised host, a failed fluconazole trial, or a story that suggests a virus or a pill rather than Candida. A failed trial is information. If fluconazole doesn't help a patient with a plausible Candida story, the likeliest explanation is that it isn't Candida, or isn't only Candida, and the next step is a scope, not a higher dose or a second antifungal.
The CD4 thresholds in HIV follow the same logic. Candida appears once the count falls under two hundred, and cytomegalovirus once it falls under one hundred. The lower the count, the longer the list of possible infections, and the less an empiric antifungal can be trusted to cover it.
He meets the first two. His pain is bad enough that he's stopped eating, and he's on three immunosuppressants, so a virus is as likely as yeast. An empiric week of fluconazole would leave a cytomegalovirus ulcer untreated while it deepens, so he goes to endoscopy.
At endoscopy, where to biopsy follows from where each virus lives. Herpes infects the surface lining, so the ulcer edge is sampled. Cytomegalovirus infects the tissue underneath, so the ulcer base is sampled. Biopsy the wrong part and the pathologist sees only inflammation. When it isn't clear which virus is present, both the edge and the base are biopsied, and brushings are taken for Candida, so one procedure answers all three questions.
His endoscopy shows no white plaques and a single large, deep ulcer in the distal esophagus. Biopsies from the base show large cells with owl's-eye nuclear inclusions in the stromal tissue. Biopsies from the edge show no herpes changes. He has cytomegalovirus esophagitis. He starts intravenous ganciclovir and switches to oral valganciclovir once he can swallow comfortably. His transplant team also reviews his immunosuppression, because the infection reflects how suppressed he is.
The third history clue covers the remaining causes. A patient who swallowed a caustic substance is an emergency rather than an outpatient workup, managed with airway protection, endoscopic staging of the injury, and long-term follow-up for stricture. Radiation is the other cause, and the second patient shows its acute form. A fifty-eight-year-old woman in the fourth week of chemoradiation for lung cancer develops severe odynophagia and can't keep up with fluids. She has acute radiation mucositis, which develops during or shortly after thoracic or head and neck radiation, with severe odynophagia, dysphagia, and dehydration. Treatment is supportive: topical anesthetics, opioids when the pain is severe, fluids, and nutritional support until the lining heals. Months to years later, the same patient may develop a late radiation stricture from fibrosis in the radiation field, which causes progressive dysphagia to solids, not odynophagia.
The second problem looks like reflux and isn't, and the third patient shows the most common version. A twenty-four-year-old woman was told she has severe reflux. Within ten minutes of starting most meals, food comes back up into her mouth. It tastes like what she just ate, not sour, and there's no nausea or retching first; it just arrives. She either swallows it again or spits it into a napkin. It happens nearly every day and has for a year. Omeprazole twice daily for three months changed nothing. She hasn't lost weight; she's gained a little. Her endoscopy with biopsies was normal.
A failed twice-daily PPI with a normal endoscopy means the first step is to ask whether this is reflux at all, and what comes up and when answers it. Hers is recognizable food, it comes within minutes of starting to eat, it's effortless, she re-swallows it, and twice-daily omeprazole did nothing. That's the typical rumination story: a young patient who has failed high-dose twice-daily PPI, with regurgitation right after every meal of material recognizable as the food just eaten.
Two other causes of regurgitation are ruled out from the history. Vomiting, from gastroparesis or anything else, is preceded by nausea and retching, and it's forceful; hers is neither. Achalasia also brings back undigested food, but it causes dysphagia to solids and liquids from the start, and she has no dysphagia.
Two details give the diagnosis away. Timing: episodes start within ten to fifteen minutes of beginning a meal, while the food is still in the stomach and undigested, which is why it tastes like what she just ate. Weight: patients keep their weight or gain, because much of the regurgitated food is swallowed again. A patient losing weight with regurgitation needs a different explanation.
Effortless regurgitation of a recognizable meal, within minutes of eating, re-swallowed or spat out, without nausea or retching, is rumination syndrome, one of the disorders of gut-brain signaling that are diagnosed after the structural and reflux workups come back normal. The Rome four criteria define it as effortless regurgitation of recently swallowed food into the mouth, followed by re-swallowing or spitting, not preceded by retching, on most days for at least three months, with onset at least six months before diagnosis. She meets every part. Her PPI failed for the same reason her endoscopy was normal: there's no acid injury to treat.
Often the history is enough to confirm it. When it's in doubt, one test shows rumination directly: high-resolution manometry with impedance after a meal. It catches a sudden simultaneous rise in pressure in the stomach and the esophagus, called an R wave, with no retching beforehand, often with the lower sphincter pressure dropping at the same instant. Impedance shows the meal moving back up. Her referring doctor wants proof before taking her off omeprazole, so she has the study, and it shows R waves after every meal.
The mechanism is the treatment. Rumination is a learned, unintentional behavior. Shortly after she starts eating, she contracts her abdominal wall and diaphragm without realizing it. That raises the pressure in her abdomen, squeezes her stomach against a relaxed lower sphincter, and pushes the undigested meal up into her esophagus and mouth. Her esophagus, sphincter, and stomach are normal. The pressure comes from her own abdominal muscles, which is why the endoscopy was normal and acid suppression couldn't help.
So the treatment makes that squeeze impossible. She can't contract her abdominal wall while breathing slowly and deeply with her diaphragm, because the two movements are incompatible. A behavioral therapist teaches her diaphragmatic breathing, slow deep breaths that push the belly out, used during and after meals, when episodes happen. Cognitive behavioral therapy is added if that isn't enough. Acid suppression has no role, so the omeprazole stops; a young patient labeled with severe refractory reflux is otherwise headed for more tests, higher doses, and sometimes surgery, and a fundoplication wouldn't stop the abdominal squeeze. After six weeks of therapy, her episodes fall from every meal to about once a week, and she's off omeprazole.
Two more patients have the same setup, reflux-type treatment that failed with a normal workup, and each has a different gut-brain disorder. The fourth patient is a fifty-two-year-old man who has had a constant feeling of a lump in his throat for six months. It's there between meals and doesn't get worse when he swallows; eating makes it better. Food doesn't stick, and swallowing doesn't hurt. A PPI didn't help. That's globus.
Globus pharyngeus is a persistent or intermittent sensation of a lump, fullness, or tightness in the throat between meals, unchanged or improved by swallowing rather than worsened. It's a sensation, not an obstruction. Dysphagia is food that sticks during a swallow; globus is a lump that's there between swallows and eases when he eats. In most patients no single physical cause is found. A hypertonic upper sphincter, post-nasal drip, or reflux sensitizing the larynx contributes in some, but the main driver is heightened perception of normal throat sensation.
So the workup is a screen rather than a search. Reflux and post-nasal drip are asked about and the throat and neck are examined. A structural lesion is looked for only if there are alarm features: true dysphagia, weight loss, hoarseness, a neck mass, or a smoking history. He has none. The treatment is reassurance, explained clearly, because globus is benign and intervention beyond reassurance rarely improves it. A PPI isn't the default answer.
The fifth patient is a forty-five-year-old woman with recurrent squeezing chest pain. Cardiology found nothing: a normal stress test and normal coronary arteries. Her endoscopy and biopsies were normal, so she doesn't have eosinophilic esophagitis. A pH study off PPI showed normal acid exposure. High-resolution manometry showed no achalasia or spasm. A PPI did nothing.
Her workup follows a fixed order, because functional chest pain is diagnosed only after the others are excluded: the heart first, because missing it is dangerous, then reflux, eosinophilic esophagitis, and a primary motility disorder. Once they're excluded, the problem is visceral hypersensitivity. The afferent nerves of her esophagus and the brain's pain pathways are set too sensitive, so ordinary events, a small reflux episode, a normal contraction, a little distension, register as pain.
Because the problem is pain processing, not acid, the treatment is a neuromodulator. A low-dose tricyclic, imipramine or amitriptyline twenty-five to fifty milligrams at bedtime, is first line. A selective serotonin reuptake inhibitor, SSRI, is the alternative if a tricyclic isn't tolerated, and hypnotherapy and cognitive behavioral therapy also work. A higher PPI dose doesn't help, because the lesion is in the pain pathway rather than the acid burden.
Two related diagnoses present as heartburn rather than chest pain, and they're separated by pH testing. Functional heartburn is heartburn with normal acid exposure and no link between symptoms and reflux events, treated with a neuromodulator. Reflux hypersensitivity is heartburn with normal acid exposure but symptoms that line up with reflux events, treated with a neuromodulator plus acid suppression, because the patient really does feel physiologic reflux events even though the acid exposure is below the threshold for reflux disease.
All three of these patients share one rule. A PPI failure is a test result: either the acid isn't suppressed enough, or acid isn't the problem. When the symptoms don't sound like reflux, immediate recognizable regurgitation, a lump between meals, or chest pain with normal acid exposure, the second explanation is far more likely. The next step is to rethink the diagnosis, not to double the dose.
Here are the traps the boards set in these patients.
In odynophagia, treating it like dysphagia misses that pain means mucosal injury. Missing the pill history misses a diagnosis the history alone can make. Giving empiric fluconazole to a transplant recipient with severe pain leaves a viral ulcer untreated. Biopsying the base of a herpes ulcer or the edge of a cytomegalovirus ulcer samples cells without the virus. Forgetting brushings for Candida, or assuming one infection when an immunocompromised patient can have more than one, leaves part of the disease untreated. And confusing acute radiation mucositis, which hurts, with a late radiation stricture, which makes food stick, sends the patient down the wrong path.
In apparent refractory reflux, doubling the PPI in a patient whose regurgitation is effortless, immediate, and recognizable treats acid that isn't the problem. Calling rumination vomiting or gastroparesis ignores the absent nausea and retching and the stable weight. And missing achalasia when regurgitation comes with dysphagia to solids and liquids overlooks a motility disorder.
In the other gut-brain disorders, ordering imaging for globus without alarm features is unnecessary. Calling chest pain functional before the heart, reflux, eosinophilic esophagitis, and motility have been excluded risks missing a real cause. And treating functional chest pain with more acid suppression instead of a neuromodulator treats the wrong pathway.
So, start to finish.
Painful swallowing means injury to the lining. Ask about a pill, immune status, and prior reflux, caustic ingestion, or radiation. Pill esophagitis follows a specific pill taken with too little water, with kissing ulcers at the aortic arch; stop the drug, give sucralfate, and teach upright dosing with water.
Candida gives white plaques and responds to fluconazole, which can be tried empirically in a mild, low-risk patient. Go straight to endoscopy when the pain is severe, the host is immunocompromised, fluconazole has failed, or the story suggests a virus or a pill. Biopsy herpes at the ulcer edge and cytomegalovirus at the ulcer base; treat herpes with acyclovir and cytomegalovirus with ganciclovir, then valganciclovir. Acute radiation mucositis is treated supportively.
When a patient who seems to have reflux fails a PPI and the workup is normal, reconsider the diagnosis before raising the dose. Effortless regurgitation of recognizable food within minutes of eating is rumination; manometry with impedance shows the R wave, and diaphragmatic breathing treats it.
A lump in the throat between meals that improves with swallowing is globus; screen for alarm features and reassure. Chest pain after negative cardiac, reflux, eosinophilic, and motility workups is functional chest pain, treated with a low-dose tricyclic.
Episode three is dysphagia after surgery on the stomach or esophagus: what each antireflux and weight-loss operation does to the swallow, and why the first test after a fundoplication is barium rather than an endoscope.