Pathogenesis

In the traditional view, reflux esophagitis is caused by reflux of gastric or duodenal fluid into the esophagus. The injurious effect of refluxed gastric acid, bile, pepsin, and duodenal contents on the normal protective mucosal barriers causes injury to the esophageal mucosa.^4–6 With ongoing reflux injury, surface esophageal cells die, triggering both an inflammatory response (infiltration of neutrophils) and a proliferative response (basal cell and papillary hyperplasia). The cause of reflux is multifactorial. Contributing factors include the presence of a hiatal hernia, a defective or weak lower esophageal sphincter (LES), impaired esophageal peristalsis with transient LES relaxation, delayed gastric emptying, decreased salivary gland secretions, increased gastric acid production, and bile reflux.^2,3,7–9 Although the data have been somewhat conflicting, there is some evidence to support an association among high body mass index, the presence of hiatal hernia, and GERD.^10–13 Likewise, there is evidence to suggest that Helicobacter pylori infection may actually protect against the development of GERD and its complications.¹⁴ Specifically, corpus gastritis is associated with decreased acid secretion, and GERD is less common in patients with severe corpus gastritis.^7,15 When acid secretion returns to normal after the eradication of H. pylori, there is an increased risk of developing GERD.^16,17

Recent studies of the pathogenesis of GERD in animal models ^15,18,19 and in vitro indicate that exposing esophageal squamous epithelial cells to acids and bile salts alone can cause epithelial cells to secrete inflammatory cytokines (i.e., interleukin 8 [IL8] and IL1β) that cause inflammatory cells (T-lymphocytes and neutrophils) to migrate into the epithelium. This finding has led to the new hypothesis that it is the inflammatory response, not the direct effect of acid, that is the initial factor responsible for damaging esophageal mucosa. Histologic examination in animal models has demonstrated that infiltration of the submucosa by lymphocytes is an early manifestation of inflammation. Neutrophils involve the mucosal surface later in the course of disease progression. These findings provide further support for an alternative concept for the development of reflux esophagitis in humans.¹⁵

Pathology

On endoscopic examination, patients with erosive esophagitis reveal erosions, ulcers, strictures, or some combination of these. However, as much as 50% to 60% of all symptomatic patients with objective evidence of GERD have normal mucosa or only mild hyperemia at endoscopy.²⁰ Furthermore, histologically inflamed esophageal mucosa (esophagitis) may appear normal endoscopically; and conversely, hyperemia does not necessarily indicate the presence of esophagitis microscopically. Because of these endoscopic/pathologic discrepancies, biopsies are always warranted in symptomatic patients to document the presence of tissue injury and to exclude other entities such as infections, BE, and other preneoplastic or inflammatory alterations. This approach is particularly important if empiric reflux therapy has failed.

Biopsies of grossly visible lesions in GERD characteristically reveal evidence of “active” esophagitis, a nonspecific injury pattern that can result from a variety of causes. Pinch biopsies obtained through standard endoscopes are often not adequate for evaluation of early histologic changes resulting from reflux because they usually do not include the entire thickness of the mucosa and are difficult to orient.^4,20 Endoscopes with a large-caliber biopsy channel and jumbo biopsy forceps should be used to facilitate accurate histologic diagnosis. Use of jumbo forceps does not increase the rate of complications related to endoscopy²¹; rather, it greatly improves the quality of the histologic specimens.

Histologically, reflux changes are typically distributed over the distal 8 to 10 cm of the esophagus in a patchy fashion, and multiple biopsies are often necessary to consistently demonstrate histologic abnormalities. ²² However, changes are typically worse in the distal esophagus and decrease in intensity more proximally. Because biopsy specimens from the lower 1 to 2 cm of the esophagus, even in asymptomatic subjects, often reveal evidence of mild squamous hyperplasia,^22,23 diagnostic biopsy specimens should be obtained usually more than 2.0 cm above the level of the gastroesophageal junction (GEJ) in order to diagnose esophagitis reliably. Furthermore, because of a high level of discordance between endoscopic and histologic findings, it is recommended that all symptomatic patients undergo biopsy, regardless of the presence or absence of endoscopic abnormalities. Esophageal biopsies are considered more useful for establishing a diagnosis of GERD in infants than in adults.²⁴

Reflux esophagitis produces a characteristic, although nonspecific, tissue injury pattern. Features of untreated active esophagitis include basal cell hyperplasia, elongation of the lamina propria papillae into more than two thirds of the thickness of the mucosa, epithelial cell necrosis, increased intraepithelial inflammation (including eosinophils, neutrophils, and lymphocytes), lack of surface maturation (nucleated cells at surface of epithelium), distended pale squamous “balloon” cells, intercellular edema (acantholysis), and, in severe cases, surface erosions or ulcerations (Figs. 14.1 to 14.3). Intercellular edema or dilated intercellular spaces reflecting increased paracellular permeability may be a useful marker of early injury in the absence of endoscopic evidence of injury.^25,26

Other Microscopic Features of GERD

Dilation and congestion of lamina propria capillaries are additional characteristic features of reflux esophagitis,⁴⁴ but this finding may also occur in specimens from normal controls, albeit in a mild fashion, possibly as a traumatic biopsy artefact.⁴⁵ Other features that may be seen in GERD include ballooning degeneration of squamous cells, intercellular edema (acantholysis) that causes minor separation of individual squamous cells, multinucleation of squamous cells, increased mitoses, and decreased surface maturation. The presence of multinucleated (regenerative) cells may mimic herpetic esophagitis (Fig. 14.4). Multinucleated regenerative squamous cells are distinguished from virally infected cells by the lack of characteristic nuclear inclusions and by their prominence at the base of the mucosa and adjacent to ulcers, rather than in the surface cells, which is characteristic of herpes.⁴⁶

Differential Diagnosis of GERD

The differential diagnosis of GERD includes infectious esophagitis; pill esophagitis; esophagitis caused by ingestion of corrosive agents such as lye; chemotherapy- and radiation-induced esophagitis; primary EOE; systemic diseases such as collagen vascular disease, Crohn’s disease, Stevens-Johnson syndrome, various bullous diseases, lichen planus, and graft-versus-host disease; and trauma. Many of these conditions share overlapping morphologic features with GERD. Therefore, an accurate diagnosis of “reflux” esophagitis requires correlation with the patient’s clinical, endoscopic, manometric, and histologic data. In the absence of clinical information, a diagnosis of reflux esophagitis cannot be established on the basis of biopsy findings alone. The histologic features are essentially nonspecific. In fact, in this scenario, a “top line” diagnosis of “active esophagitis consistent with reflux” rather than “reflux esophagitis” should be used.

Reactive Squamous Hyperplasia versus Dysplasia

On occasion, marked reactive (pseudoepitheliomatous) hyperplasia related to reflux esophagitis may resemble squamous dysplasia histologically (Fig. 14.5). The most helpful distinguishing feature is the presence of cytoarchitectural uniformity in cases of hyperplasia, compared with cytoarchitectural pleomorphism in cases of dysplasia or carcinoma.

The mucosal architecture in hyperplastic lesions retains its uniformity, showing elongation of papillae that extend to roughly equal depths within the deep lamina propria and are usually of similar width. In contrast, dysplastic squamous epithelium typically displays architectural distortion with absent, sharply angulated, or markedly irregular papillae in terms of their length and width (Table 14.1; see also Chapter 24 for details).

Cytologically, hyperplastic squamous epithelial cells are uniform and do not show loss of polarity or overlapping nuclei. The nuclei may be uniformly enlarged; however, they have smooth nuclear membranes, open chromatin, often prominent nucleoli, and increased mitoses but no atypical mitoses. Dysplastic squamous epithelial cells are typically more pleomorphic in size and shape, are more hyperchromatic, have irregular nuclear contours, and reveal nuclear overlapping and loss of polarity (see Fig. 14.5).

Reactive Changes in Ulcers versus Dysplasia or Carcinoma

Granulation tissue within the base of erosions or ulcers may exhibit large atypical endothelial cells and fibroblasts^47,48 (Fig. 14.6). They are usually distributed in a scattered fashion within otherwise typical granulation tissue. These cells do not form solid clusters of cells and have a normal, or even decreased, nucleus-to-cytoplasm (N : C) ratio. In contrast, carcinoma usually demonstrates groups or sheets of cohesive cells with overlapping nuclei and an increased N : C ratio. In difficult cases, immunohistochemistry for cytokeratins can be helpful in differentiating true carcinoma (positive) from reactive “pseudosarcomatous” alterations of the stromal cells (negative).

Rarely, the inflammatory exudate within the ulcer or erosion may contain abundant activated and atypical lymphocytes that can simulate lymphoma (Fig. 14.7). In general, these cells are benign when confined to the surface exudate. The infiltrate should raise concern for a lymphoma if it involves the underlying tissue in a dense, confluent, and homogeneous manner.

Natural History and Treatment of GERD

The natural history of GERD in the general population remains uncertain because of the widespread use of acid inhibitors. GERD has three phenotypic presentations: nonerosive reflux disease (NERD), erosive esophagitis, and BE. Most patients with GERD (50% to 70%) have normal mucosa on endoscopy and therefore are diagnosed with NERD.^49–51 Some researchers have proposed that NERD is a discrete entity because of its unique physiologic characteristics, which include a more competent antireflux barrier.⁵² However, most authorities believe that GERD is a progressive disease that starts with NERD and progresses with time, to erosive esophagitis or BE or both, in selected high-risk individuals. Some data suggest that progression from NERD to erosive esophagitis occurs in as much as 30% of patients annually,^53–56 but it is unknown whether NERD can progress directly to BE without an erosive phase.⁵⁷ Between 1% and 22% of patients with erosive esophagitis progress to a more severe form of disease, whereas regression to a less severe form of disease occurs in 6% to 42% of patients, depending on whether acid inhibitors have been used.^58–60 Erosive esophagitis is a risk factor for BE.^61,62 Barrett’s esophagus develops in between 1% to 13% of patients with erosive esophagitis annually.⁶³ In a recent prospective, follow-up endoscopic study⁶⁴ to evaluate the risk of BE in a Swedish general population (the Kalixanda study database),⁶⁵ the incidence of BE was 9.9 per 1000 person-years, and the prevalence of BE in this GERD cohort increased from 3% to 8% during a 5-year follow-up period. In fact, progression of GERD to BE in the general population may be higher than previously recognized.^55,66

The vast majority of patients with GERD have a recurrent but nonprogressive form of disease that is controlled adequately with acid inhibitor therapy.

Current medical forms of therapy, such as proton pump inhibitors (PPIs) and other acid inhibitors, are highly effective at relieving symptoms. However, progressive disease develops in some patients despite ongoing therapy.⁶⁷ In one study, as much as 82% of patients who showed healing of their esophagitis with omeprazole relapsed within 6 months after cessation of therapy.⁶⁷ In a 10-year follow-up study of 101 patients with reflux esophagitis, significant morbidity related to GERD developed in almost 75%, showing quality of life scores significantly lower than those of a non–GERD control population.⁶⁸ Surgical treatment is considered an option for patients with chronic GERD who are not responsive to medical therapy or have a defective LES.^69,70

Herpes esophagitis occurs primarily in immunosuppressed patients. Common causes of immunosuppression include prior chemotherapy, solid organ and bone marrow transplantation, and acquired immunodeficiency syndrome (AIDS). Herpes esophagitis also may occur in otherwise healthy young adults with normal immune function, who usually have a self-limited infection that resolves within 1 to 2 weeks.^72–74 Symptoms of herpetic esophagitis include odynophagia, dysphagia, epigastric pain, fever, and upper GI bleeding, but some patients are asymptomatic.⁷⁵ Coexistent herpes labialis and oropharyngeal ulcers are seen in approximately one fourth of patients.^71,74 Herpetic ulcers in the esophagus may serve as a portal of entry for other pathogens and can be associated with herpetic pneumonitis.⁷² Endoscopically, herpetic ulcers are typically shallow, sharply punched-out lesions and are often surrounded by relatively normal-appearing mucosa.

Pathology

Herpes simplex or varicella-zoster virus infects esophageal squamous epithelium. Accordingly, the characteristic inclusion bodies are limited to the squamous epithelial cells, typically accentuated at the superficial lateral margin of ulcers and erosions. In fact, biopsy specimens obtained distant from the immediate edge of the ulcer lesion may not be diagnostic. Microscopic diagnostic criteria include the presence of Cowdry A intranuclear viral inclusion bodies, ground-glass nuclei, nuclear molding, multinucleated giant cells, and ballooning degeneration of infected cells (Fig. 14.8).^72,73 Cowdry A inclusions are eosinophilic to amphophilic round bodies separated by a clear zone from a thickened nuclear membrane. Ground-glass nuclei have a smooth, homogeneous chromatin pattern with a pale basophilic quality. Multinucleated giant cell changes in squamous epithelial cells may occur in reflux esophagitis and should not be confused with the cytopathic effects of herpesvirus infection. Reactive cells have prominent nucleoli and perinucleolar clearing, but nuclear inclusion bodies are not present. Large numbers of mononuclear cells, primarily aggregates of macrophages with convoluted nuclei, in the surface exudate adjacent to the infected epithelium have been noted as a characteristic finding in herpetic ulcers and should make the pathologist suspect herpesvirus infection.⁷⁶ However, similar cells may be observed in nonherpetic ulcers throughout the GI tract in the absence of herpesvirus infection, so this finding is not specific. Herpes simplex type I is the most common cause of herpetic esophagitis, but on morphologic grounds, this type cannot be distinguished from herpes simplex type II or varicella-zoster. Immunohistochemical staining and in situ hybridization, if clinically indicated, can help distinguish these three viral species.

Treatment