‘Pill Oesophagitis’

‘Pill oesophagitis’ occurs secondary to caustic injury to the oesophagus, caused by the retention of a pill within the oesophagus. This is often associated with failure to consume adequate amounts of liquid with tablet or capsule medications, and with taking these medications in the supine position before bedtime.³ Elderly patients and female patients are most often affected.⁴

Most patients present with odynophagia, retrosternal pain, and dysphagia. The usual sites of involvement are the mid oesophagus, at the level of the aortic arch (at 22–24 cm), and, in patients with left atrial enlargement, the distal oesophagus (at 30–35 cm).³ Endoscopic findings include erythema, mucosal denudation, discrete ulcers, or erosions and strictures.³ Histological evaluation often shows the usual features of oesophagitis: acute inflammation, erosions or ulcers, and granulation tissue. Polarisable crystalline material may be an important clue to the diagnosis of ‘pill oesophagitis’, although it is not present in all cases.³

The most commonly reported agents causing pill oesophagitis include non-steroidal anti-inflammatory drugs (NSAIDs; the most common offending drug by far),⁴ antibiotics (particularly clindamycin, doxycycline, and tetracycline), potassium chloride, iron supplements, and ascorbic acid and alendronate (a bisphosphonate).⁵

Sloughing Oesophagitis

Sloughing oesophagitis (see also Chapter 11) describes corrosive injury of the oesophagus and may be caused by the ingestion of chemicals, such as lye (also known as sodium hydroxide or caustic soda).⁶ This type of injury is common in children,^{7, 8} who typically ingest caustic substances accidentally. A subset of all caustic ingestions represents suicide attempts. These patients are at lifelong risk for squamous cell carcinoma of the oesophagus.⁹ The characteristic feature of caustic injury is superficial mucosal necrosis that produces white plaques or membranes on endoscopic examination (‘oesophagitis desiccans superficialis’). This type of injury affects the mid and distal oesophagus most profoundly.¹⁰

Microscopically, the superficial epithelium shows coagulative necrosis, with separation of the superficial epithelium from the basal layer (Figure 11.7). An associated intraepithelial inflammatory infiltrate is usually seen.¹⁰

Table 5.1 summarises patterns of oesophageal pathology that may occur as a result of medications.

Iron

Patients may receive oral iron supplements, most commonly in the form of ferrous sulphate tablets, for the treatment of iron deficiency anaemia. There is no doubt that iron has the capacity to cause corrosive injury in the upper GI tract,¹⁰ and many patients on oral iron complain of upper GI symptoms including dyspepsia and nausea.¹¹ The endoscopic appearance of iron-related pathology varies from superficial erosions to frank ulceration.¹¹

Iron pill–associated upper GI pathology is increasingly recognised by pathologists (Fact Sheet 5.1; see also Chapter 13).¹¹ Mucosal erosion or ulceration may occur in the oesophagus and/or the stomach. Biopsy specimens typically display iron deposition, characterised by luminal crystalline iron (detectable on routine haematoxylin and eosin [H&E] stains), adjacent to the surface epithelium or admixed with a fibrino-inflammatory exudate (Figure 13.18).¹¹ Crystalline iron deposition can also be apparent in the lamina propria, either covered by an intact epithelium, subjacent to small superficial erosions, or admixed with granulation tissue. Iron may be deposited in lamina propria histiocytes and rarely in glandular epithelial cells. The presence of intracellular iron deposits (particularly in epithelial cells of deep glands), in the absence of a significant inflammatory component, is more suggestive of haemosiderosis or haemochromatosis than iron pill–associated mucosal injury.¹²

Occasionally, an exuberant proliferation of reactive fibroblasts and regenerative epithelial changes is present near oesophageal ulcers that contain iron crystalline material. These changes can be so striking that they raise the differential of a malignant process.¹⁰ Pseudoepitheliomatous (‘pseudocarcinomatous’) epithelial hyperplasia may also appear in areas of reparative squamous oesophageal mucosa, causing initial alarm.¹⁰

Bisphosphonates

Bisphosphonates (e.g. alendronate sodium) are effective in the treatment of osteoporosis, Paget’s disease, and hypercalcaemia of malignancy, in order to prevent osteoclast-mediated bone reabsorption.¹³ Ingestion of alendronate sodium (Fosamax) and related medications is associated with both oesophagitis and oesophageal ulcers.^{14, 15} Patients may develop erosive/ulcerative oesophagitis during alendronate therapy.⁵

Oesophageal biopsies of the ulcer site show a characteristic inflammatory exudate with granulation tissue. Polarisable, crystalline, foreign material may be present (pale yellow crystals), and multinucleated giant cells within the inflammatory exudate may be found near the crystalline foreign material.¹ Adjacent squamous epithelium typically shows active inflammation with reactive/regenerative epithelial changes, characterised by epithelial cells with enlarged, hyperchromatic nuclei.

Reactive/Chemical Gastropathy

Reactive/chemical gastropathy/gastritis represents the second most common diagnosis made on gastric biopsies, after Helicobacter pylori–associated gastritis.¹⁶ It is regarded as a non-specific response to a variety of gastric irritants,¹⁷ the most common of which are NSAIDs, bile reflux, and alcohol.

The histological features of reactive/chemical gastropathy include (Fact Sheet 5.2) hyperplastic foveolar epithelium with a ‘corkscrew’ appearance (tortuosity of gastric pits), surface epithelial degeneration with cuboidal change of the foveolar glandular cells (nuclear enlargement and hyperchromasia), lamina propria oedema, dilated capillaries, and vascular congestion in the lamina propria, a paucity of inflammatory cells, and smooth muscle splaying in the lamina propria with vertical orientation of the muscle fibres (see also Chapter 13).¹⁸ Foci of pseudopyloric or intestinal metaplasia are common and may reflect ulcer repair.¹⁸

Ulceration or Perforation

Numerous drugs, most commonly NSAIDs, can cause erosions or ulcers in the stomach.¹⁹ Other drugs associated with GI ulcers include corticosteroids, alendronate, doxycycline, colchicine, chemotherapy, iron, and Kayexalate.

Mucosal Deposits

OsmoPrep is a tablet form of bowel preparation.²⁰ The active ingredient in OsmoPrep is sodium phosphate, which functions as an osmotic laxative. It is a colon preparatory agent that is used infrequently, primarily because of the associated risk of acute phosphate nephropathy.²⁰ Patients may be prescribed this form of colon preparation if they cannot tolerate the large volume of chalky fluid that constitutes the more commonly prescribed liquid sodium phosphate agent.²⁰

OsmoPrep-associated gastritis is characterised by marked reactive epithelial changes, associated with purple to black inorganic deposits in the superficial lamina propria of gastric antral biopsies.²⁰ The gastric mucosa associated with the deposits typically demonstrates prominent mucin loss and nuclear hyperchromasia, without significantly increased inflammation, producing changes compatible with reactive gastropathy. The deposits have been described as irregular in contour, of varying sizes (typically <100 mm), and generally having the appearance of crushed pill fragments. Histochemical special stains show that the deposits are Perls iron stain negative and von Kossa stain positive.²⁰ The histological differential diagnosis includes mucosal calcinosis and iron pill gastritis.²⁰

Table 5.2 summarises patterns of gastric pathology that may occur as a result of medications.

Colchicine

Mucosal changes are observed only when this alkaloid reaches toxic levels (typically in patients with renal failure or hepatic failure).²¹ The histological changes reflect, in part, the inhibition of tubulin polymerisation.¹⁷ The gastric epithelium frequently shows nuclear pseudo-stratification and loss of polarity.²¹ Numerous mitotic figures arrested in metaphase are seen, with the chromosomes often arranged as ‘ring’ mitoses.²¹ Apoptoses can also be prominent, and apoptotic bodies are typically located either in the proliferative region of the gastric crypt or of the gland neck.¹⁷

Aluminium-Containing Antacids or Sucralfate

Gastric mucosal calcinosis is seen most frequently in orthotopic transplant patients (varied organs) and in chronic renal failure patients who have been prescribed either aluminium-containing antacids or sucralfate.¹⁷ Gastric mucosal calcinosis describes the presence of small, deeply pink, and partially calcified refractile crystals, that characteristically reside beneath the surface epithelium of the gastric antrum.¹⁷ Typically, some degree of foveolar hyperplasia and mucosal oedema are also present.²²

Radiation

Gastric mucosal injury may be seen following radiation therapy for upper abdominal neoplasia or in bone marrow transplant (BMT) recipients.¹⁷ Radiation gastritis is diagnosed less commonly than radiation enteritis (see also Chapter 3).

Early changes (8–10 days post-irradiation) are characterised histologically by nuclear karyorrhexis and cytoplasmic eosinophilia of the gastric pit epithelium (Fact Sheet 5.3). During the subsequent days, mucosal oedema and congestion develop, together with submucosal collagen bundle swelling, fibrin deposition, and telangiectasia. Inflammation is usually insignificant. Characteristic radiation-induced nuclear atypia follows. If damage is extensive, ulceration and haemorrhage may occur. Possible late radiation effects include endothelial proliferation and fibrinoid necrosis of blood vessel walls. Recovery usually begins during the third week after radiation injury and is complete within 2–3 months.¹⁷

Yttrium-90–Microsphere Selective Internal Radiation

Selective internal radiation therapy (SIRT) (see also Chapter 3) with biocompatible resin-based yttrium-90 (Y-90)-emitting microspheres, administered via hepatic artery branches, is a method used to deliver internal radiation therapy selectively to inoperable/unresectable primary and secondary (particularly colorectal liver metastases) hepatic malignancies.^{23, 24} As the overwhelming majority of the tumour blood supply is derived from the hepatic artery (while the normal parenchyma is largely supplied by the portal vein), this method allows selective delivery of effective doses of radiation to liver lesions without compromising normal liver parenchyma.²⁵Thus, it is a form of loco-regional radiation therapy.

However, selective internal radiation therapy has the potential to cause extrahepatic adverse effects if the microspheres are inadvertently delivered to arteries supplying the stomach or duodenum or other organs.²⁵ The reported incidence of GI complications of this therapy varies from 3% to 24%.²⁶ Radiation ulcers are a recognised complication, situated mostly in the gastric antrum, pylorus, and duodenum^{23, 26} and occur as a consequence of indirect irradiation or direct deposition of Y-90 microspheres.²³

The presence of round, purple microspheres on histological examination of biopsy specimens is diagnostic of direct deposition of Y-90 microspheres (Figures 3.14 and 6.1B).^{23, 27} The microspheres are 20–30 μm in diameter and this size allows them to lodge preferentially in the microvasculature of the tumour as well as in the microvasculature of the GI tract, leading to direct radiation toxicity.^{23, 26, 27} Even small quantities of microspheres caught in the gastric and/or duodenal capillary bed may lead to ulceration, bleeding, and perforation.²⁴ Other changes that may be appreciated histologically include apoptosis, epithelial flattening, glandular cystic dilatation, nuclear atypia, capillary ectasia, and prominent endothelial cells.²⁵ The lack of lamellations differentiates the spheres from psammoma bodies.²³

Duodenal Villous Atrophy

Some therapeutic agents can cause villous atrophy in addition to significant intraepithelial lymphocytosis, with or without epithelial damage (Fact Sheet 5.5).

Fact Sheet 5.5 Medications Causing Small Intestinal Villous Atrophy

Mycophenolate

Mycophenolate (MPA) is an immunomodulator used to prevent rejection in solid organ transplant, including heart, kidney, and liver. Two formulations are available, mycophenolate mofetil (MMF) and mycophenolate sodium (MPS),⁴⁵ commonly marketed under the trade names CellCept (Roche Pharmaceuticals) and Myfortic (Novartis), respectively. MMF is absorbed in the stomach while MPS is absorbed in the small intestine.⁴⁶ MPA inhibits the enzyme inosine monophosphate dehydrogenase and prevents the conversion of inosine monophosphate into guanosine monophosphate, decreasing purine synthesis and arresting the cell cycle. This in turn affects the production of T and B lymphocytes, resulting in immunosuppression.^{47, 48} Common GI side effects of MPA (affecting up to 45% of patients⁴⁹) include nausea, vomiting, diarrhoea and abdominal pain, and range from mild (intermittent nausea or diarrhoea) to life-threatening (colonic necrosis or perforation) complications.⁵⁰

Histological abnormalities in the upper GI tract of patients treated with MPA reflect local mucosal irritation, similar to that induced by NSAIDs,⁵¹ and include ulcerative oesophagitis, reactive gastropathy-type changes, and duodenal ulcers.⁴⁴ Coeliac disease–type changes in duodenal biopsies may also occur in patients treated with MPA. The first histological report of duodenal MPA-associated injury appeared in 1998, when villous blunting and crypt hyperplasia were described in a renal transplant patient on MMF therapy.⁵²

Increased epithelial cell apoptosis has been identified as a hallmark feature for the diagnosis of MPA-related colitis.⁵³ Striking apoptosis in duodenal and colonic biopsies from patients on MPA therapy has been described.^{43, 44, 48, 53} Other morphological features independently associated with MMF include the presence and quantity of lamina propria eosinophils and endocrine cell aggregates, and the presence and quantity of apoptotic microabscesses, hypereosinophilic degenerated crypts, and crypt distortion^{53, 54} (Figure 5.1; Fact Sheet 5.8).

(A) A medium-power view of colonic mucosa with markedly increased numbers of eosinophils in the lamina propria, and a crypt abscess.

(B) A high-power view of colonic mucosa with increased numbers of eosinophils in the lamina propria and an apoptotic microabscess with cryptitis.

Figure 5.1 Mycophenolate-associated colitis.

Ipilimumab

Ipilimubab, a fully human monoclonal antibody directed against cytotoxic T-lymphocyte antigen-4, has been shown to offer durable objective therapeutic responses in patients with metastatic malignant melanoma and renal cell carcinoma.⁶³ However, this regimen is associated with numerous immune-mediated toxicities, including enterocolitis, hepatitis, nephritis, dermatitis, hypophysitis, and uveitis.^63–67 Major toxicity appears to affect the GI tract most frequently, occurring in up to 20% of patients receiving ipilimumab.⁶⁸ A 5% mortality rate has been reported in patients who develop ipilimumab-associated colitis, with a significant risk of colonic perforation.^{64, 69} Endoscopic findings in patients suffering from severe ipilimumab-associated GI toxicity can be variable, ranging from normal mucosa to diffusely erythematous and ulcerated mucosa.⁷⁰

Histologically, ipilimumab-associated enteritis demonstrates features in ileal and duodenal biopsies resembling coeliac disease or autoimmune enteropathy, namely villous blunting, an increase in intraepithelial lymphocytes, and an expansion of the lamina propria by a lymphoplasmacytic infiltrate (Fact Sheet 5.10). Increased epithelial apoptosis is often a distinctive feature (Figure 5.3).

(A) A low-power view of colonic mucosa with increased cellularity of the lamina propria and increased epithelial apoptosis.

(B) A high-power view of colonic mucosa with a plasma cell infiltrate in the lamina propria and several apoptotic bodies.

Figure 5.3 Ipilimumab-associated colitis.

Histological findings of ipilimumab-associated colitis include an inflammatory infiltrate in the lamina propria comprising neutrophils and plasma cells, and focal neutrophilic and lymphocytic cryptitis (Fact Sheet 5.10).⁶⁶ Foci of crypt abscesses, glandular destruction, and erosions of the mucosal surface are usually evident, with occasional ulceration. These inflammatory changes are often diffuse throughout the colon.

The colitis associated with ipilimumab has histological features similar to those of colorectal graft-versus-host disease (GvHD) as well as to those of inflammatory bowel disease (IBD) (see also Chapter 22, including acute and chronic inflammatory changes and patchy areas of inflammation (skip lesions).⁶⁶ GvHD is usually histologically distinct, being characterised by prominent epithelial cell apoptosis and glandular destruction. These features are less prominent in ipilimumab-associated colitis. Unlike Crohn’s disease or ulcerative colitis, ipilimumab-related colitis involves the descending colon more than the sigmoid colon, ascending colon or rectum⁷¹ and does not usually demonstrate features of chronicity.⁶⁶

Crystal Deposition: Kayexalate

Sodium polystyrene sulphonate (Kayexalate; SPS) is used routinely to treat hyperkalaemia (by binding and excreting potassium through the GI tract), mainly in patients with end-stage renal disease (ESRD) but also in those with hyperkalaemia resulting from other diseases.¹ SPS may cause ischaemic necrotising lesions in any segment of the intestine (especially in the small intestine and colon) that can appear within hours and up to 11 days after intake.⁷²

Histological features of cation-exchange resin-induced injury include transmural necrosis and ulcerations containing resin crystals that can be present in the lumen, or on the serosal surface if perforation occurs.⁷² Kayexalate crystals have narrow, rectangular ‘fish scales’, are violet/lightly basophilic on H&E and are refractile but not polarisable (Figure 5.4).⁷³ The crystals stain red with periodic acid–Schiff (PAS) and Ziehl–Neelsen (ZN) stains, magenta with periodic acid–Schiff diastase (PASD) and display a characteristic crystalline mosaic pattern (Table 5.5).⁷⁴

Figure 5.4 Kayexalate crystals have narrow, rectangular ‘fish scales’, are lightly basophilic, and are refractile but not polarisable.

(Courtesy of Dr Laura Lamps, University of Michigan, USA.)

Focal Active Colitis

Focal active colitis (FAC) is the term used to describe the isolated finding of neutrophilic crypt injury (focal acute inflammation of the crypt epithelium) of the colorectal mucosa, diagnosed in colorectal biopsies, in the absence of any other significant microscopic abnormality (Fact Sheet 5.11).^{75, 76} This descriptive term encompasses a spectrum of histological changes, ranging from a single focus of cryptitis or a single crypt abscess to multiple discrete foci of cryptitis or crypt abscesses within a series of colorectal biopsies.⁷⁵

Fact Sheet 5.11 Medications Causing Focal Active Colitis

Studies on adult populations with FAC have suggested that the majority of cases of FAC are due to drugs (especially NSAIDs) and self-limiting colitis⁄infection.^{75, 77, 78} Drugs that have often been implicated include antibiotics, PPIs, steroids, NSAIDs(mefenamic acid, diclofenac, naproxen), and immunosuppressive medication.^{78, 79} However, drugs that have been reported to cause FAC are numerous, and also include the following: ipilimumab, bisoprolol, furosemide, amiodarone, thyroxine, simvastatin, Asacol, fluoxetine, Sanomigran, pirprofen, carbamazepine, oral contraceptive steroids, laxatives (such as bisphosphonate enemas and bisacodyl) and oral bowel preparation.⁷⁹ Please see Table 5.4.

Ischaemic Colitis

More than 90% of patients with non-iatrogenic causes of ischaemic colitis are in their seventh decade of life or older. Occasionally, ischaemic colitis occurs in younger patients as a result of comorbidities or medications. There is a well-established association between ischaemic colitis and illicit drugs, such as cocaine,^80–82 and prescription drugs, such as ergotamine,⁸³ oestrogens,^84–86 progesterones,⁸⁷ and sodium polystyrene. Psychotropics, particularly neuroleptics, gold, and beta-blockers, are other possible risk factors for ischaemic colitis.^88–90 Phenothiazine, clozapine, and tricyclic antidepressants are the most frequently reported anti-psychotics thought to cause ischaemic colitis.^91–94

A probable correlation between a wide range of drugs (including alosetron, interferon-a, digitalis, dopamine, epinephrine or norepinephrine, amphetamines, methylsergide, NSAIDs, pseudo-ephedrine, vasopressin, barbiturates, cyclosporine, danazol, diuretics, flutamide, glycerin enema, and phospho-soda solution) and ischaemic colitis has been reported.⁸⁸ In addition, isolated case reports have described ischaemic colitis as a possible consequence of cyclo-oxygenase 2 (COX2) inhibitors, triptans, chemotherapy, tegaserod, clinasetron, phentermine, voglibose, RhinAdvil, vinorelbine, combination therapy with peginterferon and ribavirin (described in more detail in the text that follows), oral laxatives, and barium enema.⁸⁸

Eosinophilic Colitis

A consensus on the diagnostic criteria for eosinophilic colitis does not currently exist, although many pathologists use a diagnostic threshold of 20 eosinophils per high-power field (hpf).⁹⁵ However, normal values for tissue eosinophils vary widely between different segments of the colon, ranging from <10 eosinophils per hpf in the rectum to>30 eosinophils per hpf in the caecum.⁹⁶ Therefore, the location of the biopsy is critically important for the interpretation of findings.⁹⁷ Drug-induced eosinophilic colitis has been described in response to a variety of drugs, including NSAIDs,⁹⁸ clozapine,^99–101 carbamazepine,¹⁰² rifampicin,¹⁰³ tacrolimus,¹⁰⁴ and gold.¹⁰⁵

Microscopic-Type Colitis

Microscopic colitis is a common cause of chronic diarrhoea, especially in the elderly.¹⁰⁶ A diagnosis of microscopic colitis (lymphocytic or collagenous colitis) is based upon histological abnormalities of colonic mucosa in patients with chronic diarrhoea and macroscopically normal or near normal colonic mucosa (see Chapter 20).¹⁰⁷ Although the pathophysiology of microscopic colitis is not well understood it is probably multifactorial and related to an abnormal immune reaction to luminal antigens in predisposed hosts.¹⁰⁸ Recent research has highlighted the potential role of drugs in inducing microscopic colitis.¹⁰⁷ Identifying such drugs is important, considering the associated disease burden.

NSAIDs (can be associated with collagenous colitis or lymphocytic colitis),^{109, 110} PPIs (e.g. lansoprazole; can lead to either lymphocytic [stronger association] or collagenous colitis),^{109, 111} and serotonin reuptake inhibitors (SSRIs; e.g. sertraline, associated with lymphocytic colitis) are among the medications that are very likely to induce microscopic colitis.^112–115 However, an ever-increasing list of drugs is associated with microscopic colitis and includes 3-hydroxy-3-methylglutaryl-coenzymeHMG-CoA reductase inhibitors (statins, e.g. simvastatin), ticlopidine, ranitidine, flutamide, acarbose, carbamazepine, penicillin, and aspirin.^{111, 116–120} Drug-induced microscopic colitis is an important condition to recognise, as removal of the causative agent can result in resolution of diarrhoea and avoid the cost and potential side effects of specific pharmacotherapy for microscopic colitis.^{114, 121}