The Stomach Looks Suspicious, But Is It Pernicious?

This article has an accompanying continuing medical education activity, also eligible for MOC credit, on page e9. Learning Objective: Upon completion of this CME activity, successful learners will be able to 1) identify endoscopic features of gastric intestinal metaplasia and apply current guidelines for management; and 2) identify clinical manifestations and complications of autoimmune gastritis.

This case was discussed among a multidisciplinary panel consisting of two gastroenterologists (SCS, DL), an expert gastrointestinal pathologist (MBP), and an endocrinologist (SB). The discussion, which occurred via e-mail conversations and a video conference, is summarized here.

Case History

A 65-year-old Mandarin-speaking gentleman presents with abdominal discomfort for the past 5 months. He denies nausea, vomiting, early satiety, weight loss, or any change in his appetite or bowel movements. His medical history is significant for coronary artery disease with cardiac stent placement in 2018, hypothyroidism, hypertension, hyperlipidemia, and vitamin B₁₂ deficiency (receiving B₁₂ injections). He denies any prior knowledge or treatment of Helicobacter pylori infection, but his daughter tested positive and was successfully eradicated. He recently underwent a computed tomography scan of the abdomen and pelvis with intravenous and oral contrast a few months ago for evaluation of mild unintentional weight loss that has since resolved. The computed tomography scan was unremarkable. His family history includes breast cancer in his mother. His social history includes a former 20-pack-year smoking history (quit in 1980), 4 glasses of wine per week, and no other substance use. He is a first-generation immigrant from China.

On physical examination, he is a well-appearing, Mandarin-speaking man in no acute distress or discomfort. Abdominal examination demonstrated normoactive bowel sounds and a nontender, nondistended abdomen. A review of the laboratory studies (reference values in parentheses) that were available at the time of the office visit demonstrated hemoglobin 10 g/dL (13.7–17.5 g/dL), mean corpuscular volume 92 μm³ (79.0–95.0 μm³), vitamin B₁₂ 205 pg/mL (211–946 pg/mL), which was before initiating supplementation, ferritin 10 ng/mL (30–400 ng/mL), iron saturation 19%, total iron-binding capacity 366 μg/dL (148–506 μg/dL), and thyroid-stimulating hormone (TSH) 1.39 μIU/mL (0.27–4.20 μIU/mL) on levothyroxine.

He underwent esophagogastroduodenoscopy (EGD) 2 months before his visit today, which demonstrated “general loss of mucosal folds suggestive of gastric atrophy.” There was a 1-cm gastric polyp with superficial erosion that was biopsied, and a 2-mm nodule along the lesser curvature of the gastric body that was removed with cold forceps, retrieved, and placed in a separate jar. Additionally, nontargeted gastric biopsies were obtained from the antrum and body in a random fashion and combined in the same jar. The random biopsies demonstrated gastric oxyntic mucosa with oxyntic gland atrophy and dropout, mild chronic inflammation, neuroendocrine hyperplasia, and intestinal metaplasia, and antral-appearing mucosa with mild chronic inflammation, but without intestinal metaplasia. The gastric polyp pathology demonstrated a hyperplastic polyp with focal erosion and maturing granulation tissue, without intestinal metaplasia or dysplasia. The gastric nodule demonstrated well-differentiated neuroendocrine tumor (NET) with Ki-67 proliferation index of <3%. No H. pylori was identified on routine stain or immunohistochemistry stain. Colonoscopy was performed at the same time and, other than hemorrhoids, was normal. Based on the EGD and biopsy findings, he had additional laboratory studies performed. His antiparietal cell antibody was 54.2 U (0.0–24.9 U), intrinsic factor (IF) antibody was negative, and gastrin level was elevated at 1154 pg/mL (0.00–100 pg/mL). H. pylori fecal antigen testing was negative.

Question: What is the best next step in this patient’s management?

1. Treat for H. pylori infection.

2. Schedule for upper EGD with mapping biopsies.

3. No further management is indicated.

4. Start proton pump inhibitor therapy.

Multidisciplinary Case Discussion

SCS: This patient’s clinical presentation and investigative workup including EGD, histology, and laboratory studies are suggestive of a diagnosis of autoimmune gastritis (AIG). Dan, does this patient fit the “classic” demographic profile for AIG that we see portrayed in questions on the GI boards? What are your thoughts?

DL: Shailja, thank you for selecting this great case for GI grand rounds. AIG is a chronic inflammatory condition of the stomach caused by autoimmune destruction of parietal cells of the gastric corpus, which leads to progressive atrophy of the gastric oxyntic mucosa with classic sparing of the gastric antrum. It remains an underrecognized condition because most patients with AIG lack specific symptoms, such as the patient in this case, and may go undiagnosed for many years. The estimated prevalence of AIG ranges from 0.5 to 2.0% in the general population, but the range varies based on the population and heterogenous study designs.¹ The median age of AIG diagnosis ranges from 50 to 69 in different studies and most studies reported a female predominance, with female to male ratio of 2–3:1.^2–4 AIG has been reported across different racial and ethnic groups worldwide; therefore, race and ethnicity are not particularly useful factors to rule out AIG.

So, although his age of 65 years still falls into the common age range for AIG, it is less typical for males to be diagnosed with AIG compared with females. Notwithstanding, several findings from his workup point support a diagnosis of AIG. First, his endoscopy showed loss of mucosal folds, a hallmark endoscopic feature of corpus atrophy. Other typical endoscopic findings (not mentioned in the endoscopic report of this case) of gastric mucosal atrophy include pallor of the mucosa (Figure 1) and increased visibility of submucosal blood vessels due to the destruction of the oxyntic mucosa. Another clue for AIG in this patient is the pathology confirming oxyntic gland atrophy, mild chronic inflammation, neuroendocrine (enterochromaffin-like [ECL] cell) hyperplasia, and a small NET. ECL cell hyperplasia is caused by chronic hypergastrinemia induced by hypochlorhydria or achlorhydria due to parietal cell loss.¹ NET is the long-term consequence of ECL cell hyperplasia. However, ECL cell hyperplasia and NET are not 100% specific for AIG and may be seen in severe H. pylori-associated pangastric atrophy as well. The presence of autoantibodies such as parietal cell antibody (PCA) or IF antibody (IFA) is helpful, but not necessary, for the diagnosis of AIG. PCA is more sensitive (sensitivity of approximately 70%–80%), but less specific because it can also be positive in H. pylori associated corpus inflammation. IFA is less sensitive (30%–60% sensitivity) but highly specific (>95% specificity).^2,4,5 This patient has an elevated PCA but negative IFA, which is suggestive but still is not a “slam-dunk” for an AIG diagnosis. His elevated gastrin level indicates hypochlorhydria or achlorhydria due to parietal cell loss, which resulted in a compensatory hypersecretion of gastrin from the antrum (assuming also that he is not on a proton pump inhibitor (PPI), which can also increase gastrin level through a similar mechanism). He also has hypothyroidism and B₁₂ deficiency, and both are associated with AIG due to autoimmune thyroid disease and lack of IF for B₁₂ absorption, respectively. His mean corpuscular volume was within the normal range, but may represent a combination of microcytosis due to his iron deficiency and macrocytosis from the B₁₂ deficiency. It should be noted that iron deficiency is not uncommon in patients with AIG¹ and is also related to achlorhydria and hypochlorhydria, because the acidic environment facilitates iron absorption. Taken together, his workup does support a diagnosis of AIG.

A very interesting aspect of this case is that the patient most likely also has H. pylori-associated atrophic gastritis (HpAG) based on the finding of chronic gastritis on antral biopsy and a large hyperplastic polyp (which is suggestive of, but certainly not diagnostic of, chronic H. pylori infection). From an epidemiologic standpoint, atrophic gastritis caused by H. pylori infection is significantly more common than AIG. As a first-generation immigrant from China who is in his mid-60s, this patient has a very high likelihood of H. pylori exposure. Interestingly, H. pylori was negative on both gastric biopsies and stool antigen test, despite his denying ever being treated, which may suggest that he either inadvertently cleared the bacteria through prior antibiotic consumption for other indications or his test results represented a false negative; spontaneous clearance of H. pylori is certainly possible, but rare. If his daughter has lived with him in the same house, her H. pylori positivity would support the assumption that he has been infected too.

As a gastroenterologist, it is essential to remember that histology remains indispensable for the diagnosis of both AIG and HpAG. Specifically, biopsies should be obtained from both the antrum and corpus and placed in separate specimen bottles. This process allows pathologists to assess the extent and severity of inflammation and any preneoplastic changes (eg, mucosal atrophy and intestinal metaplasia) in the antrum and corpus separately, which is important for both diagnosis (eg, HpAG vs AIG) and risk stratification with respect to the risk of future neoplasia. For example, normal antral biopsies along with the presence of corpus atrophy or ECL-cell hyperplasia (with or without NETs) sufficiently rules out HpAG and supports the diagnosis of AIG—essentially, corpus atrophy with antral sparing. Patients with extensive or severe gastric atrophy or gastric intestinal metaplasia (GIM) based on gastric biopsies have a significantly increased risk for noncardia gastric adenocarcinoma.^6–9

For this patient, because all the initial random biopsy specimens were placed into the same bottle, it is reasonable to bring him back for a repeat endoscopy for mapping biopsies of gastric corpus and antrum to further assess the topographic presence and severity of atrophic gastritis and GIM. This examination would also serve to rule out subtle prevalent neoplasia that might have been missed on his initial examination, particularly because elements such as adequacy of mucosal visualization, time spent examining the stomach, and the use of narrow-band imaging or other image enhancement are not mentioned.^10,11 Additionally, I would favor endoscopic removal of the 1cm hyperplastic polyp at the time of repeat endoscopy, given its malignant potential, although the small risk of complications from resection such as bleeding should be discussed.

Figure 1.

Representative endoscopic images of autoimmune gastritis involving gastric corpus with an overlap of Helicobacter pylori-associated atrophic gastritis involving the antrum. (A, B) Gastric atrophy of the corpus characterized by the loss of normal rugal folds and mucosal pallor. (C, D) Autoimmune gastritis with biopsy-proven areas (white arrows) of enterochromaffin-like cell hyperplasia and focal neuroendocrine tumor (type 1) on near-focus white-light endoscopy (C) and narrow-band imaging (D). (E, F) H. pylori-associated atrophic mucosa involving the antrum and incisura with pallor of the mucosa.

SCS: Dan, these are all great points. As you mentioned, it is curious that this patient’s H. pylori stool antigen and histology did not demonstrate active H. pylori infection and he denied prior H. pylori treatment. He was not on PPI or other medications that could decrease the sensitivity of these testing modalities. But, as you just pointed out, he is certainly at high risk for having H. pylori exposure. Is there any value in testing H. pylori serology? We know that there are limitations of H. pylori serology testing including differences in test performance based on population prevalence of H. pylori and suboptimal specificity in some populations, namely, individuals who are older and with atrophic gastritis¹²—similar to this patient. H. pylori serology testing is also no longer offered at many centers because of these suboptimal test characteristics. What are your thoughts?

DL: Shailja, this is a great question, and all accurate points. As we know, current professional recommendations such as ones from the American College of Gastroenterology, Houston Consensus and Maastricht VI/Florence consensus report all recommend using nonserological methods such as urea breath testing or stool antigen or gastric biopsies to confirm active H. pylori infection before prescribing eradication therapy.^13–15 However, I have to say, that because a positive serology may last for decades after H. pylori infection, serology is still useful in certain situations, such as assessing a prior H. pylori infection to inform risk stratification, particularly for patients with AIG (which I will discuss soon) or in situations such as acute bleeding peptic ulcer disease where patients cannot reasonably discontinue PPI treatment for a sufficiently long time (ie, 2 weeks). To this end, the recent Maastricht VI/Florence consensus suggests serological tests can be of particular value in situations such as bleeding peptic ulcers, gastric mucosa-associated lymphoid tissue lymphoma, gastric adenocarcinoma, gastric atrophy, and recent use of antibiotics or PPI that cannot be discontinued for ≥2 weeks.¹⁵

Another particular reason to obtain serology is that the knowledge of H. pylori infection may help to inform his future risk of gastric adenocarcinoma. AIG is known to be associated with an increased risk of both gastric adenocarcinoma and type 1 NET.^1,4 However, recent evidence suggests the risk of gastric adenocarcinoma in AIG is not elevated unless there is previous or current H. pylori infection. In a prospective cohort study by Rugge et al,¹⁶ 211 patients with AIG were followed and H. pylori infection was rigorously ruled out by serological, histological, and molecular methods. After a mean follow-up of 7.5 years and 10,541 person-years, no incident invasive gastric adenocarcinoma cases were detected in this cohort, although progressive neuroendocrine proliferation was observed during the follow-up. These findings need to be further validated in different populations and with a longer follow-up time, but this study provides provocative data to support the lack of substantially increased risk of gastric adenocarcinoma in patients with AIG in the absence of H. pylori infection. The study also supports the importance of H. pylori screening and eradication in patients with AIG. But, as you mentioned, age is one consideration, and it is possible that his serology may be falsely negative now if he cleared the bacteria many years ago. So, I am not sure I would feel all that reassured by a negative serological test when considering his very high pretest probability for H. pylori exposure based on his age, immigration from China, his daughter testing positive, and also considering that the inflammatory changes noted on antral biopsies and intestinal metaplasia and atrophy on corpus biopsies could certainly be related to chronic H. pylori infection. This case highlights a challenge in real-world practices when it may be difficult to convincingly rule out H. pylori with the available test methods in patients with AIG. To this end, I would rely on good-quality gastric biopsies to guide my practice (additional support for answer choice B). If the histology supports H. pylori-associated changes, I would still consider surveillance for these patients for both adenocarcinoma and type 1 NET, even if the H. pylori testing result was negative.

SCS: Blanca, I am going to turn it over to you now for the pathology perspective. Many of us stress the importance of separating the biopsies antrum and corpus, but it is not always done in routine practice. Can you comment from the pathology standpoint why separating the biopsies is recommended? Would you be able to make the diagnosis of AIG without the biopsies separated?

MBP: Thanks, Shailja, for inviting me to provide the pathology perspective on this very interesting case of potential AIG and HpAG overlap. Separating biopsies into antral/incisura and corpus compartments facilitates the assessment of the preservation of the native mucosa as well as the inflammatory changes in each compartment in routine hematoxylin and eosin staining. In addition, when multiple biopsies are obtained from each compartment, it facilitates the proper orientation of the tissues during the paraffin-embedding process, which is needed for a proper assessment of the severity of the atrophic changes. In any case, pathologists should be aware of the histologic features of AIG and carefully evaluate the tissue independent of specimen labeling. Histopathological features that strongly suggest a diagnosis of AIG include a moderate or marked loss of parietal and chief cells replaced by pseudopyloric and intestinal metaplasia in a background of chronic inflammation in the corpus (Figure 2A), and a relatively normal antral mucosa. HpAG may show similar changes in the corpus, but will typically show antral involvement, with inflammatory and atrophic changes. Having said that, the presence of inflammatory or atrophic changes in the antral mucosa does not discard the possibility of AIG, because both conditions may coexist, as seems to be the case here. If the anatomical location of the tissues is uncertain, an immunohistochemical stain for gastrin will discriminate between antral/transitional mucosa and pseudopyloric metaplasia—that is, gastrin will be present and stain in the former, but is absent in the latter. However, such an approach may not be helpful in poorly oriented biopsies with only foveolar mucosa represented in the sections, or in biopsies in which the gastric mucosa is completely replaced by intestinal metaplasia.

Figure 2.

Histopathologic features of AIG in the gastric corpus. (A) Complete absence of parietal and chief cells replaced by pseudopyloric and intestinal (arrows) metaplasia, in a background of chronic inflammation (stain: hematoxylin and eosin; original magnification ×100). (B) Synaptophysin stain highlighting linear (arrows) and micronodular (arrowheads) ECL cell hyperplasia (original magnification ×200).

SCS: Thank you, Blanca! Could you help us to understand the finding of neuroendocrine hyperplasia? What differentiates this from NET? Can you see these findings in patients with chronic H. pylori infection but not AIG? (ie, is this a sine qua non of AIG?)

MBP: The progressive loss of parietal cells causes a state of hypochlorhydria (Figure 2A). This stimulates the feedback loop that regulates acid secretion, causing the antral G cells to increase gastrin production in an effort to stimulate the secretion of acid by parietal cells. At the same time, the continuous gastrin production stimulates the ECL cells of the gastric corpus, leading to ECL cell hyperplasia. Linear ECL cell hyperplasia consists of groups of ≥5 contiguous ECL cells lining the gastric glands (in a linear fashion) (Figure 2B), and micronodular hyperplasia consists of minute aggregates of ≥5 ECL cells.¹⁷ As proliferation of ECL cells continues, NETs (type 1) may arise, as occurred in this patient. NETs are defined by the presence of invasion regardless of size,¹⁸ and, in the setting of gastric atrophy, they are typically low grade, defined by a mitotic rate of <2 mitoses/2 mm² and a Ki67 labeling index of <3%,¹⁹ as the one observed in this patient. The neuroendocrine differentiation may be confirmed using an immunohistochemical stain for synaptophysin or chromogranin A. Because the oxyntic atrophy extensively affects the gastric corpus, multiple NETs may be observed.¹⁷ Although infrequent, ECL cell hyperplasia and NETs may occur in the setting of H. pylori infection without serological evidence of AIG.²⁰ Unfortunately, there is no way to definitively differentiate findings consistent with AIG from so-called burnt-out H. pylori gastritis (ie, no evidence of current active H. pylori infection) and clinicians should correlate the histologic findings with the clinical scenario and laboratory tests (eg, PCA, IF).

SCS: Dan, can you tell us your approach to someone with gastric NET?

DL: Fortunately, most of the NETs in patients with AIG are type 1 NETs which represent 80%–90% of all gastric NETs. These lesions typically appear as tiny to small (<10 mm) nodules at the gastric corpus or fundus on endoscopy, as in our patient here (Figure 1). Type 1 gastric NETs tend to be well-differentiated, with an indolent course. Small type 1 gastric NETs <1 cm are amenable to endoscopic resection, with an excellent prognosis. For larger NETs, the estimated rate of metastasis is <10% for those ≤2 cm but approaches 20% in NETs >2 cm.¹ The optimal interval for endoscopic surveillance for patients with AIG with or without NETs has not been well-defined. For AIG without NET but already with ECL cell hyperplasia, it may be reasonable to repeat an EGD in 3 years. For those with small NETs, after endoscopic resection of NETs, a repeat EGD can be considered in 1–3 years, depending on the size and burden of the tumors. For gastric NETs >1–2 cm, consider endoscopic ultrasound examination, multiphasic computed tomography scan, or magnetic resonance imaging to assess the depth of tumor invasion and local or distant metastasis to help guide further management.²¹ Surgical resection is recommended for NETs >2 cm, deeper invasion (beyond submucosa), or with lymph node metastasis. In contrast, type 2 gastric NETs are caused by hypergastrinemia related to pancreatic or duodenal gastrinoma (Zollinger–Ellison syndrome) and typically are multifocal, but fortunately also with an indolent course. Type 3 gastric NETs are sporadic unifocal neoplasms without any association with either AIG or gastrinoma and have more aggressive clinical behavior.²¹

DL: Shailja, how would you approach surveillance in this gentleman with likely overlap of AIG/HpAG, complicated by GIM? Is there any role for screening family members?

SCS: Dan, thanks for these important questions. This case is certainly not straightforward from a surveillance standpoint, but I will walk through some things that I am thinking about. First, I agree that an EGD with mapping biopsies is indicated because his prior EGD, although recent, was not sufficient for risk stratification and we also do not know the quality of the EGD. The purpose of the endoscopic surveillance examination in someone with gastric preneoplasia is to (1) detect neoplasia and (2) risk stratify.¹¹ This patient is at high risk for future gastric neoplasia as well as gastric neoplasia that could have been missed on his most recent EGD. His risk factors include age, first-generation immigrant from a country where the gastric cancer incidence is high, smoking history, corpus intestinal metaplasia, and high likelihood of prior H. pylori infection (or potentially current infection if his testing was falsely negative). It is still not definitive that he has AIG and pernicious anemia; nevertheless, some older data suggest that the risk of gastric cancer in patients with pernicious anemia is highest within the first year of diagnosis.^22,23 Certain important details are missing from his first endoscopy that would be helpful for risk stratification, including the severity of underlying atrophy and GIM, as well as histologic subtype of GIM. Gastric neoplasia, including early stage gastric adenocarcinoma, may be subtle and the endoscopic miss rate based on numerous studies, including meta-analyses is approximately 10%.^24,25 Therefore, the importance of performing a high-quality endoscopy cannot be overemphasized, which I would like to elaborate on a bit more.

A high-quality examination of the stomach includes the following elements: (1) using high-definition white light endoscopy (over standard definition) and ideally with image-enhancing technology (such as narrow band imaging); (2) cleansing the gastric mucosa to achieve adequate mucosal visualization by clearing debris, fluid, and bubbles; (3) using sufficient insufflation to distend the gastric lumen and flatten the gastric folds; (4) ensuring sufficient observation time to visualize all stations of the gastric compartment; (5) performing photodocumentation of the entire gastric mucosa; and (6) obtaining biopsies according to the updated Sydney system biopsy protocol with additional targeted biopsies of abnormal lesions. The optimal gastric examination time is unclear. One study from Singapore demonstrated that endoscopists with total examination times of ≥7 minutes were 2.5 times more likely to detect high-risk gastric lesions compared with those with shorter examination times,²⁶ which is consistent with other literature.^27,28 This is analogous to colonoscopy withdrawal time and the impact on adenoma detection rate. Adequate photodocumentation is another measure to ensure a high-quality examination. The recent World Endoscopy Organization position statement proposed a protocol for complete photodocumentation of the upper gastrointestinal tract using 28 photos, including 22 gastric photos documenting from the gastroesophageal junction to the pylorus²⁹, which parallels the “systematic screening protocol for the stomach” originally developed in Japan. When balancing feasibility in US clinical practice without compromising diagnostic yield and quality, we suggest ≥6 photographs from the following gastric stations such that the full gastric mucosa is documented: cardia/fundus and incisura on retroflexed view, corpus lesser curvature, greater curvature, and antrum/pylorus on forward view. Areas suspicious for preneoplasia or neoplasia should be photographed separately. Last, it is important to follow a systematic biopsy protocol to increase the yield and accuracy in detecting preneoplastic and neoplastic changes and risk stratifying the patient. For US gastroenterologists, a practical approach is to follow the updated Sydney biopsy protocol, which requires 2 biopsies taken from the antrum (one each from the lesser curvature and greater curvature), 1 biopsy from the incisura, and 2 biopsies from the gastric corpus (1 each from the lesser curvature and greater curvature) and placed in separate specimen bottles.^10,11 Additional targeted biopsies should be obtained from visible mucosal abnormalities and placed in separate specimen bottles.

With respect to the surveillance of atrophic gastritis and GIM, the American Gastroenterological Association guidelines for GIM management generally align with the European Society of Gastrointestinal Endoscopy and British Society of Gastroenterology.^1,7–9 Surveillance for the purposes of gastric neoplasia detection is also recommended in patients with AIG, but the guidance, including recommended intervals, is less clear.¹ For this patient, if he is confirmed to have only complete-type GIM of the corpus (which is typical of AIG), with mild corpus atrophic gastritis in all biopsies and mild atrophic antral gastritis without more advanced changes, I would recommend a surveillance endoscopy in 3 years (Figure 3). The histologic subtype of GIM can be determined using standard hematoxylin and eosin staining in the vast majority of cases without the need for additional special stains, assuming an appropriately cut and well-oriented biopsy specimen.³⁰ Complete-type GIM resembles small intestinal epithelium, whereas incomplete-type GIM resembles colonic epithelium. Further subtyping of incomplete GIM (eg, type II vs type III) requires special stains but this is typically reserved for research purposes and is not routine clinical practice. Compared with the complete-type GIM, incomplete-type GIM is associated with a significantly higher risk of gastric cancer (generally 4- to 5-fold higher) based on several studies and meta-analyses of these studies. Additional risk factors to recognize include family history of gastric cancer in a first-degree relative, persistent H. pylori infection (despite eradication attempts), extent and severity of preneoplastic changes, and possibly smoking, among others. Notably, race, ethnicity, and country of origin are risk factors for gastric preneoplasia, but once atrophic gastritis and GIM are diagnosed, these are not independently associated with the risk of progression to gastric adenocarcinoma.

Family clustering has been reported in patients with AIG,³¹ similar to other autoimmune diseases. In 1 study, 16.5% of patients with AIG had ≥1 first-degree relative with AIG.³² However, data are insufficient to recommend screening asymptomatic family members of patients with AIG. However, it is reasonable, and suggested, to screen for H. pylori infection in family members who live in the same household as an individual with a history of H. pylori infection.¹³

Last, it is relevant to also discuss nonendoscopic management apart from H. pylori testing and eradication in patients diagnosed with gastric preneoplasia. Specifically, data consistently show an increased risk of gastric cancer associated with diets high in salt, smoked and processed meats, and red meats, whereas fresh fruits and vegetables, especially those high in vitamin C, are associated with a reduced risk of gastric cancer. Smoking is a major risk factor for gastric cancer with 1 meta-analysis of 42 observational studies demonstrating that current cigarette smokers had a 53% higher risk of gastric cancer compared to nonsmokers, and this risk seems to be dose and duration dependent.³³ As such, I always counsel patients regarding dietary modifications and behavioral modifications, such as smoking cessation, as appropriate.

Figure 3.

Algorithm for endoscopic surveillance in patients with moderate to severe atrophic gastritis and/or gastric intestinal metaplasia (GIM). All patients with atrophic gastritis and/or GIM should be tested for Helicobacter pylori using nonserological methods, treated if positive, and undergo repeat nonserological testing to confirm eradication ≥4 weeks after treatment completion (off of proton pump inhibitors). The endoscopic surveillance examination should be performed with a high-definition white light endoscopic system and image enhancement technologies (such as narrow-band imaging), and adequate mucosal visualization should be achieved via appropriate cleansing and insufflation. This algorithm assumes shared patient–provider decision-making and also assumes that patients are medically appropriate for endoscopic surveillance. AG, atrophic gastritis; EGD, esophagogastroduodenoscopy.

SCS: Schafer, how do you approach a diagnosis of pernicious anemia? Does this patient meet criteria for the diagnosis? Are there other diagnostic tests you would recommend?

SB: Normally, vitamin B₁₂ dissociates from food proteins in the acidic environment of the stomach before forming a complex with IF, a glycoprotein produced by the gastric parietal cells. The B₁₂–IF complex is later absorbed via mucosal receptors in the distal ileum. Conditions or medications that interfere with this process decrease vitamin B₁₂ absorption, which can impair myelin sheath formation and lead to adverse central and peripheral nervous system manifestations. In pernicious anemia, there is autoimmune destruction of the parietal cells, which leads to achlorhydria or hypochlorhydria, and production of anti-IF antibodies that block B₁₂–IF complex formation. This leads to impaired absorption of B₁₂ and causes megaloblastic anemia. There are no universally accepted diagnostic criteria for pernicious anemia, although the workup may include a complete blood count with peripheral smear; iron panel, B₁₂, folic acid, methylmalonic acid, and homocysteine; PCA and IFA; endoscopy; and (rarely) bone marrow biopsy.

The combination of low serum B₁₂ with positive IFA is highly specific for the diagnosis of pernicious anemia, although not highly sensitive. Functional IFAs are estimated to be present in only 50%–70% of individuals with pernicious anemia,³⁴ and B₁₂ deficiency can be present even when serum B₁₂ levels are within normal limits (especially if methylmalonic acid is elevated). PCAs, which are not believed to directly contribute to the pathogenesis of pernicious anemia, are relatively nonspecific and should not be considered sufficient to make the diagnosis.

Our patient is negative for IFA and does not have macrocytosis, possibly owing to concomitant iron-deficiency anemia, as Dan mentioned. However, the combination of vitamin B₁₂ deficiency, positive PCA, and the endoscopic and histopathologic findings could be consistent with pernicious anemia. Other causes of vitamin B₁₂ deficiency include inadequate dietary intake, medications (eg, long-term PPI or metformin therapy), gastric surgery, and ileal disease affecting absorption. Treatment with high-dose vitamin B₁₂ supplementation is appropriate to minimize the risk of neurologic damage.

SCS: Thank you for that great explanation. Does the diagnosis of hypothyroidism in this elderly gentleman make the diagnosis of AIG more likely in your mind? How do you approach hypothyroidism in patients with AIG? Do you screen for other autoimmune conditions?

SB: Autoimmune conditions tend to travel together, so the presence of hypothyroidism (presumed to be Hashimoto’s thyroiditis) does increase my suspicion for AIG. Approximately 50% of individuals with AIG have another autoimmune disease, and some have multiple.^2,4,35 The autoimmune conditions that are most frequently associated with AIG are autoimmune thyroid disease (usually Hashimoto’s and rarely Graves’ disease), type 1 diabetes, Addison disease, and vitiligo. Routine screening for type 1 diabetes and Addison disease is not recommended, but given the potential dangers of these diseases, providers should have a low threshold for testing in the presence of potential signs or symptoms. Hypothyroidism owing to Hashimoto’s is by far the most common comorbid autoimmune disease, and screening with a TSH level is advised. There are no consensus guidelines for the frequency of rescreening if the initial TSH is normal; however, it is reasonable to repeat the TSH for screening at least every 5 years or for diagnostic purposes if there are new symptoms (Figure 4).

Figure 4.

Algorithm for the clinical management of atrophic gastritis (AG), including autoimmune gastritis. AIG, autoimmune gastritis; IF, intrinsic factor; PCA, parietal cell antibody; NET, neuroendocrine tumor. Adapted from the American Gastroenterological Association (AGA) Clinical Practice Update on Atrophic Gastritis (Shah SC, et al. Gastroenterology 2021; 161(4):1325–1332).

SCS: Thank you all for this fantastic discussion. As this case clearly illustrates, it can be difficult to make a diagnosis of AIG, particularly given the overlapping features of H. pylori-associated mucosal changes. As we have discussed, AIG is rare and pernicious anemia is a late-stage manifestation, whereas HpAG is substantially more common. One thing is clear—this patient is at high risk for gastric cancer and does warrant consideration for a high-quality endoscopy to rule out dysplasia or malignancy and more adequately assess his risk. This process should always be preceded by shared decision-making that involves a careful, individualized discussion of risks versus benefits.