NSAID-induced pyloric stenosis leading to oesophageal intramucosal dissection

Abstract

We describe a rare case of a 75-year-old woman with significant non-steroidal anti-inflammatory drug (NSAID) use who presented with haematemesis. Upper endoscopy revealed a large (9 cm) intramucosal dissection of the oesophagus without extension into the gastro-oesophageal junction and a severely narrowed pylorus. We postulate that she developed pyloric stenosis due to peptic ulcer disease from chronic NSAID use. This then led to gastro-oesophageal reflux. Undigested pills in the refluxate had contacted oesophageal mucosa, leading to pill-induced oesophageal injury. This, along with vomiting, is postulated to have led to the oesophageal intramucosal dissection. She improved with conservative medical management with a clear liquid diet and proton pump inhibitors, and a follow-up upper endoscopy 1 week later showed recovery of the previously seen intramucosal dissection.

Background

Pill-induced oesophagitis, eosinophilic oesophagitis and ulcerated reflux oesophagitis can be difficult to differentiate due to their similar clinical features. This case highlights the importance of careful history taking in identifying the possible culprit of a life-threatening condition. This is also an excellent example of the various pathological effects of non-steroidal anti-inflammatory drugs (NSAIDs) on the gastrointestinal tract, including peptic ulcer-related pyloric stenosis and oesophageal ulcer complicated by intramucosal dissection. This case serves as a reminder to be cautious prior to prescribing NSAIDs to an at-risk group.

Case presentation

Our patient was a 75-year-old woman who never sought care from a primary care physician. She admitted to a long-standing history of regurgitation, intermittent acid reflux and waterbrash. In addition to this, she described symptoms of intermittent postprandial inability to tolerate secretions, which would resolve spontaneously. She suffered from frequent migraine headaches and took 3–4 tablets of an over-the-counter NSAID, which contained acetaminophen, aspirin and caffeine every day for nearly 20 years. She was never formally diagnosed with peptic ulcer disease but took significant amounts of calcium bicarbonate for relief of dyspepsia. Her surgical history includes hysterectomy, tonsillectomy and no prior upper endoscopy or colonoscopy. She had no history of chronic alcohol use, cigarette smoking or illicit drug, toxin or caustic agent ingestion.

She presented to an outside facility with acute onset of vomiting and haematemesis, which began several hours prior to admission. She was eating dinner and had spontaneous vomiting of food, followed by ∼400 mL of bright red blood. She denied abdominal and chest pain. Prior to this, she had no odynophagia, no significant weight loss and no history, haematemesis or melaena.

On initial examination, her blood pressure was 80/40 mm Hg, heart rate 87 bpm and afebrile. Cardiovascular and respiratory examinations were unremarkable. There was no crepitus on palpation of the chest or neck. The abdomen was soft; bowel sounds were present, but there was no palpable mass.

At the outside hospital, she was emergently intubated for airway protection. She then underwent an upper endoscopy which revealed a 9 cm×2 cm intramucosal dissection of the oesophagus and was transferred to our facility for evaluation by the cardiothoracic surgeon for possibility of perforation.

Investigations

She received 2 units of red blood cell transfusion prior to transfer from the outside facility. Her haemoglobin on arrival was 15.1 g/dL, platelet of 177 k/uL, International Normalised Ratio 1.0. Other blood parameters including liver enzymes and renal functions were within normal limits.

CT scan of the chest and abdomen with oral contrast (gastrograffin) administered via a nasogastric tube placed at the mid oesophagus revealed intraluminal contrast, with extension of the contrast into the oesopahageal wall and thickening of the distal oesophagus, as shown in figure 1. There was no extravasation of contrast or pneumomediastinum. The distal thickening of the oesophagus and gastro-oesophageal junction were better defined on MRI chest with contrast as oedema in the oesophageal wall. There was no definitive evidence of mass or haematoma in the oesophageal wall.

Figure 1.

Axial view of CT of the chest with oral contrast demonstrating extraluminal contrast adjacent to the oesophagus lumen, suggesting of intramucosal dissection.

An upper endoscopy was performed on day 2 of admission to further investigate the cause of haematemesis. There was a large oesophageal intramucosal dissection (figure 2), which extended from 27 to 36 cm beyond the incisors. The dissection had no stigmata of active bleeding and did not extend into the gastro-oesophageal junction, which was at 39 cm from the incisors. There was a large blood clot in the stomach. Intrinsic pyloric stenosis was also found (figure 3) and could not be traversed with an adult diagnostic endoscope. An ultra-thin scope was subsequently used and the examined portion of the duodenum was found to be unremarkable.

Figure 2.

Upper endoscopy showing oesophageal intramucosal dissection extending to 9 cm in length, without involvement of the gastro-oesophageal junction.

Figure 3.

Narrowed pylorus which could not be traversed with an adult diagnostic endoscope, with a background of a dark blood clot in the stomach.

Differential diagnosis

At this point, there were three main differential diagnoses for this patient's presentations. Pill-induced oesophagitis, eosinophilic oesophagitis and ulcerative reflux oesophagitis. These could all present as a complicated case of oesophageal ulcers and intramucosal dissection, resulting in haematemesis. In her case, the presence of significant narrowing of the pylorus was most likely due to benign peptic ulcerations from her chronic NSAID use. Her pyloric stenosis could also explain her long-standing history of regurgitation and reflux symptoms.

Treatment

During the endoscopy, the pylorus was also dilated using a 6–7–8 mm through-the-scope hydrostatic balloon dilator. There was no immediate complication postprocedure. The patient was managed conservatively postendoscopy. She was subsequently extubated and started on a clear liquid diet, proton pump inhibitor and sucralfate. She was also given adequate antiemetic agents and analgesics to prevent any sudden increase in intra-abdominal pressure.

Outcome and follow-up

The patient was monitored closely and discharged home on day 6 of hospitalisation with pantoprazole and sucralfate. She tolerated a clear liquid diet well and had no recurrence of symptoms. She was also told to avoid any form of NSAID. One week later, she returned for a repeat upper endoscopy which revealed significant healing of the previously seen oesophageal ulcer, now about 6 cm in length (figure 4). There were also erythematous areas on the distal oesophagus. Biopsies of the distal and proximal oesophagus were only significant for chronic inflammation without any eosinophilic infiltration. The stenosed pylorus was further dilated to 10 mm with a through-the-scope hydrostatic balloon dilator.

Figure 4.

Repeat upper endoscopy 1 week later demonstrating a well-healed oesophagus.

Discussion

Clinically discerning pill-induced oesophagitis, eosinophilic oesophagitis and ulcerative reflux oesophagitis can be unclear due to their similar presentations. Common symptoms include odynophagia, dysphagia, vomiting and melaena. Typically, most patients with pill-oesophagitis present with more abrupt symptoms.^{1 2} Pill-induced oesophagitis usually involves the middle third of the oesophagus; eosinophilic oeosophagitis has diffuse involvement of the oesophagus while reflux oesophagitis generally involves the distal oesophagus.^{1 2}

There are clinical features that may help differentiate these conditions, which are found with varying degrees of frequency. Impacted pill fragments found in the oesophagus during endoscopic examination may suggest pill-induced oesophagitis. While patients with a history of allergies and oesophageal dysfunction following intake of triggering foods, point to the diagnosis of eosinophilic oesophagitis.³ Reflux oesophagitis may also mimic eosinophilic oesophagitis both clinically and histologically; however, it commonly affects only the distal oesophagus.² The overlapping features and lack of specific markers can make it difficult to differentiate these three diseases.

We postulated that this patient's presentation is related to her chronic excessive use of NSAIDs. Benign peptic ulcer disease is the most common cause of pyloric stenosis in adults.⁴ Her habitual ingestion of NSAIDs had resulted in peptic ulcer disease that led to a cicatricial narrowing of the pylorus. The patient's pylorus could only be traversed with an ultra-thin endoscope, with a shaft diameter of 4.9 mm. Her stenosed pylorus caused a functional gastric outlet obstruction which led to recurrent regurgitation of food content, including the NSAID pills that she ingested. The pathophysiology underlying the oesophageal injury caused by NSAIDs is similar to that of peptic ulcer disease in that there is disruption of mucosal integrity and repair due to depletion of prostaglandins.⁵ In addition to this, our patient's coexistent pyloric stenosis resulted in a repetitive exposure of the oesophageal mucosa to the direct, caustic effect of NSAIDs. Although rare, pill-oesophagitis has been reported to cause intramucosal dissection on several occasions, and in severe case causing perforation.^6–8

Conservative management is appropriate in most cases of pill-induced-oesophagitis. This includes cessation of offending agents, acid suppressant agents and cytoprotective agents.

Learning points.

• Diagnosis of pill-induced oesophagitis, eosinophilic oesophagitis and ulcerative reflux oesophagitis can be difficult due to their overlapping clinicopathological manifestations.

• Non-steroidal anti-inflammatory drugs (NSAIDs) can lead to a variety spectrum of diseases, including peptic ulcer disease, pyloric stenosis and oesophageal injury.

• Prescription of NSAIDs should be avoided, especially among patients with symptomatic gastro-oesophageal reflux disease.