Fundic gland polyps, initially described in association with

CLINICAL GASTROENTEROLOGY AND HEPATOLOGY 2009;7:849 854 No Association Between Gastric Fundic Gland Polyps and Gastrointestinal Neoplasia in a Study of Over 100,000 Patients ROBERT M. GENTA,*, CHRISTOPHER M. SCHULER,* CRISTIAN I. ROBIOU,* and RICHARD H. LASH* *Division of Gastrointestinal Pathology, Caris Diagnostics, Irving; Departments of Pathology and Medicine, Veterans Affairs North Texas Health Care System, University of Texas Southwestern Medical Center, Dallas, Texas BACKGROUND & AIMS: Fundic gland polyps (FGPs), the most common type of gastric polyps, have been associated with prolonged proton pump inhibitor therapy and an increased risk of colon cancer. The presence of FGPs has been inversely correlated with Helicobacter pylori infection. We evaluated the prevalence of H pylori-associated gastritis, colonic polyps, and carcinomas in subjects with and without FGPs. METHODS: We analyzed data collected from community-based endoscopy centers in 36 states (plus Washington DC and Puerto Rico) on patients who underwent esophagogastroduodenoscopy (EGD) and colonoscopy between April 2007 and March 2008. Of the 103,385 patients who underwent EGD during this time period, gastric biopsy samples were collected from 78,801 and colonic biopsies from 26,017. Slides of samples from Helicobacter-infected FGPs and FGPs with dysplasia were reviewed. RE- SULTS: FGPs were detected in 6081 patients (67.8% women). Helicobacter infection was present in less than 0.5% patients with FGPs and 13.0% of those without FGPs (odds ratio [OR], 29.05; 95% confidence interval [CI], 20.4 41.4; P.0001). Colonic adenomas were detected in 42.3% of women with FGPs and 33.8% of those without (OR, 1.43; 95% CI, 1.26 1.63; P.001); there was no significant difference in colonic adenomas between men with and without FGPs. CONCLUSIONS: Women had a higher prevalence of FGPs. FGPs were associated with gastroesophageal reflux disease symptoms, gastric heterotopia, hyperplastic colonic polyps (only in men), and colonic adenomas (only in women, especially those over 60 years of age). The presence of FGPs was inversely correlated with H pylori infection, active gastritis, and gastric neoplasia. Fundic gland polyps, initially described in association with familial adenomatous polyposis of the colon 1,2 were later noted in patients without colonic polyps and described in detail by Elster et al in 1977. 3 These lesions are best defined as cystic dilatations of the oxyntic glands lined by variably flattened parietal and chief cells with or without mucous foveolar cells (Figure 1). To be classified as sporadic, a fundic gland polyp must occur in a patient with no known history or concurrent evidence of familial polyposis. Sporadic fundic gland polyps (FGPs) are currently the most common type of gastric polyp, representing more than 50% of all excised gastric polyps. Although there have been dissenting views, 4,5 several retrospective studies have detected a relationship between FGPs and the prolonged use of proton pump inhibitors (PPIs). 6 9 In addition, FGPs have been suspected of being negatively associated with other gastric conditions, including Helicobacter pylori gastritis 10 13 and intestinal metaplasia. 14 A potentially more disquieting association is the alleged relationship between sporadic FGPs and colonic adenomas 15,16 and even carcinomas. 17 The remarkable management implications of these findings are typified in these and other researchers suggestion that every sporadic FGP patient should undergo colonic surveillance. 14 These studies were conducted in two European countries (Germany and Italy) where both the population and the endoscopic practices differ considerably from those in the United States; their conclusions were derived from relatively small series of local patients; and the authors have unanimously emphasized the need for larger studies to confirm their findings. Therefore, we designed this study to characterize the clinicopathologic associations of FGPs in a large American population sample. Specifically, we investigated whether the reported negative association with H pylori infection could be confirmed and whether there was, in fact, a significant association with colonic polyps (both hyperplastic and adenomatous) and/or adenocarcinoma that could support the notion that patients with sporadic FGPs need colonic surveillance. Methods Study Setting This study was conducted at Caris Diagnostics, a specialized gastrointestinal laboratory receiving specimens from gastroenterologists operating in outpatient endoscopy centers in 36 states, Puerto Rico, and the District of Columbia. Biopsies are interpreted by an experienced group of gastrointestinal pathologists who share a common approach to biopsy evaluation and maintain a relative uniformity through internally standardized specimen handling, diagnostic criteria, and terminology. Specifically, our criteria for the diagnosis of fundic gland polyps include: cystic dilatation of benign oxyntic glands in an endoscopically identified gastric polyp. Smaller, noncystic dilatations of oxyntic glands are not diagnosed as fundic gland polyps and are typically considered to be related to proton pump inhibitors. Such cases are not included in this analysis. The study was approved by the Caris Diagnostics Institutional Review Board. Because data were collected entirely by Abbreviations used in this paper: CI, confidence interval; EGD, esophagogastroduodenoscopy; FGP, fundic gland polyp; GERD, gastroesophageal reflux disease; OR, odds ratio; PPI, proton pump inhibitor; VBA, Visual Basics for Applications. 2009 by the AGA Institute 1542-3565/09/$36.00 doi:10.1016/j.cgh.2009.05.015

850 GENTA ET AL CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 7, No. 8 Figure 1. (A) Fundic gland polyp (magnification 4 ): Numerous microcystic glands are noted in an irregularly expanded oxyntic mucosa. The surface epithelium is smooth, with short foveolae. (B) At higher power (magnification 20 ) one can appreciate that the microcysts are lined by budding parietal cells (arrows), chief cells, and mucous cells. There is virtually no inflammation in the lamina propria. Hematoxylin and eosin stains. reviewing existing records, no direct contact with either patients or providers was involved, and no individual information is revealed in any form, exemption from the need for informed consent from participants was granted. Sources of Data To identify the records for eligible gastric biopsy specimens, we extracted data from all cases within the Caris Diagnostics database reviewed in our Irving, Texas, location from patients who had undergone an esophagogastroduodenoscopy (EGD) within the 12-month period from April 1, 2007, to March 31, 2008. Patients who had specimens processed and diagnosed at other Caris locations (Phoenix, AZ, and Boston, MA), as well as cases received in consultation, were excluded. The database includes demographic and clinical information for each patient, as well as either a summary of the endoscopic findings or the entire endoscopic report. The site of origin of each specimen and the histopathologic report for each biopsy performed on a patient are listed in the diagnosis field, which includes biopsy site designations, procedure, and histopathologic findings for every specimen submitted for a particular patient encounter. Gastric specimens were selected by searches of the diagnoses for the term fundic gland polyp. Additional information including the presence or absence of H pylori (provided in each diagnosis) or H heilmannii, chronic active gastritis, and intestinal metaplasia were extracted by analyzing individual diagnostic lines using additional search terms and boolean logic in Visual Basic for Applications (VBA). To investigate the relationship of FGPs with colonic neoplasms, we extracted data regarding all noninflammatory and nonhamartomatous colonic polyps (hyperplastic, serrated, and adenomatous), adenocarcinomas, and carcinoids from all patients who had undergone a simultaneous colonoscopy ( same-day bidirectional endoscopy ) and, for comparison, all those who metachronously underwent a colonoscopy during the same period. Clinical indications were extracted using simple queries in Microsoft Access (Microsoft Corp, Redmond, WA). Histopathologic Review In our laboratory the presence or absence of H pylori is specifically mentioned in all final diagnostic reports. Detection is aided by a modified Giemsa stain (HP Blue, Anatech Ltd, Battle Creek, MI) that is routinely performed on all gastric biopsy specimens received. When Helicobacter organisms are not identified with HP Blue stain, but infection is nevertheless suspected (based upon the histologic findings of chronic active gastritis, lymphoid aggregates, etc), a peroxidase-conjugated polyclonal anti-h pylori immunohistochemical stain (Cell Marque, Rocklin, CA) is performed. All cases in which H pylori organisms were reported to have been detected on the gastric epithelium overlying the fundic gland polyp were reviewed and, when necessary, an immunohistochemical stain for the confirmation of Helicobacter organisms was performed. Sections from all FGPs initially diagnosed as having dysplasia were re-evaluated, and the patients information was reviewed for the possibility of polyposis syndromes. Slide reviews were performed by all authors using a multiheaded microscope; results reported here reflect their consensus. Statistical Analysis All statistical calculations were performed using SigmaStat Version 3.5 (Systat Software, Inc, Point Richmond, CA). Distributions of categorical variables were compared with an uncorrected 2 test. Means and standard deviations were calculated for continuous variables (age in years), and comparisons between groups were made by the Student t test. When normality check failed, the Mann Whitney Rank Sum Test for nonparametric data was used. Multiple correlations were calculated by logistic regression. Simple odds ratios were calculated using an Internet-based odds ratio (OR) calculator. 18 Results Prevalence of FGPs The database contained 121,564 unique patient encounters which included an EGD with biopsies; after excluding patients whose biopsies had not been diagnosed at the Caris Diagnostics Irving laboratory, those who had cytologic specimens only, and external consults, there remained a total of 103,385 unique patients (61,007 women, or 59.0%). Of these, 78,801 patients had one or more gastric biopsies, whereas the remaining 24,584 patients had biopsies from the esophagus or the duodenum, but not from the stomach; this latter group is assumed not to have gastric polyps and, specifically, not to have fundic gland polyps. Of the 103,385 patients who had an EGD, 26,017 patients also had a colonoscopy during the study period (including 16,191 patients who had a same-day bidirectional endoscopy). One or more fundic gland polyps were diagnosed in 6084 patients. Three patients (2 adult men and 1 15-year-old boy) had a previous diagnosis of familial adenomatous polyposis

August 2009 FUNDIC GLAND POLYPS, H. PYLORI, AND COLONIC NEOPLASIA 851 Table 1. Indications for EGD (75,672 Adults Who Underwent EGD Only) Indication 4475 Patients with FGP, n (%) 71,197 Patients without FGP, n (%) OR (95% CI) probability (Yates value) GERD 2083 (46.5) 26,590 (37.3) 1.46 (1.37 1.55) P.0001 Dyspepsia, epigastric pain 922 (20.6) 17,370 (24.3) 0.80 (0.75 0.87) P.001 Anemia 259 (5.8) 5315 (7.5) 0.76 (0.67 0.86) P.0001 Dysphagia, odynophagia 780 (17.4) 13,391 (18.9) 0.91 (0.84 0.98) P.05 NOTE. Patients with FGPs were significantly more likely than patients without FGPs to undergo an EGD because of complaints related to GERD, and significantly less likely to present with dyspepsia, epigastric pain, or anemia. Esophageal symptoms were virtually unrelated to the endoscopic detection of an FGP. and were excluded from this analysis, yielding a total of 6081 patients (5.9%) with presumably sporadic fundic gland polyps; 4124 (67.8%) were women (median age 59 years, range 17 92 years) and 1960 were men (median age 59 years, range 11 95 years). Thus, the OR for women to have FGP was 1.61 (95% confidence interval [CI], 1.38 1.44). Indications for Endoscopy Patients who underwent an EGD and a colonoscopy either simultaneously or during the same 12 months may have had symptoms resulting from either the upper or the lower gastrointestinal tract. To avoid this bias, we analyzed the indications for EGD only in the 75,672 adults who did not have a colonoscopy during the study period. The results are summarized in Table 1. Almost half (46.5%) of the patients found to have a fundic polyp had gastroesophageal reflux disease (GERD) as the main reason for the performance of an EGD, in contrast to 37.3% of patients who did not have an FGP (OR, 1.46; 95% CI, 1.37 1.55). Dyspepsia and epigastric pain, anemia, and to lesser extents, dysphagia and odynophagia, were all significantly less likely to be the indication for EGD in patients with FGPs than in those without FGPs. Dysplasia in Fundic Gland Polyps Epithelial changes suspicious for dysplasia were detected in FGPs from 15 patients. Slides were reviewed by the authors and the following consensus was reached: 6 patients (3 men and 3 women) had low-grade epithelial dysplasia; 10 patients (4 men and 6 women) had atypical epithelial features that did not show distinct adenomatous changes and also had evidence of surface injury or regenerative changes. No cases of high grade dysplasia or carcinoma were detected. Gastritis, Helicobacter, and Gastric Neoplasms The pathologic findings in the upper gastrointestinal tract of patients with and without FGPs are depicted in Table 2. Chronic active gastritis was present in 1.6% of patients with FGPs and in 15.1% of those without (OR, 14.98; 95% CI, 11.84 18.94); an even stronger negative association was found with Helicobacter infection, which was detected histologically in only 6 patients with FGPs (0.1%) versus 9440 (13.0%) of patients without FGPs (OR, 150.47; 95% CI, 67.55 335.16). Four of the 6 infected patients had H pylori and 2 had H heilmannii (Figure 2). The prevalence of atrophic gastritis and intestinal metaplasia was also significantly lower in patients with FGPs than in those without. No synchronous gastric adenocarcinomas, adenomas, or carcinoids were present in patients with fundic polyps; in contrast, 178 patients without FGPs had an adenocarcinoma (invasive or intramucosal), 48 had an adenoma, and 46 had a gastric carcinoid. Table 2. Associated Conditions in the Upper Gastrointestinal Tract Concurrent histologic diagnoses FGP patients, n (%) N 6081 Patients without FGP, n (%) N 72,797 Odds ratio (95% CI) P Esophagus Inlet patch 16 (0.3) 141 (0.2) 1.36 (0.81 2.29) ns Stomach CAG 95 (1.6) 11,022 (15.1) 0.09 (0.07 0.11).0001 Helicobacter 6 (0.1) 9440 (13.0) 0.006 (0.003 0.15).0001 IM 135 (2.2) 5732 (7.9) 0.16 (0.13 0.19).0001 Atrophic gastritis 11 (0.2) 637 (0.9) 0.014 (0.077 0.025).0001 Adenoma 0 63 (0.06) Adenocarcinoma (also intramucosal) 0 178 (0.14) Gastric carcinoid 0 53 (0.06) Duodenum DIL 67 (1.1) 787 (1.1) 1.02 (0.79 1.31) ns Sprue 24 (0.4) 329 (0.4) 0.87 (0.57 1.32) ns Gastric heterotopia 55 (0.9) 313 (0.4) 2.11 (1.58 2.81).0001 NOTE. Gastric heterotopia is defined as an island of oxyntic mucosa lined by gastric-type mucous epithelium; these cases do not include gastric foveolar metaplasia (patches of gastric mucous epithelium replacing small intestinal absorptive epithelium in the duodenum without underlying oxyntic glands). CAG, chronic active gastritis; DIL, duodenal intraepithelial lymphocytosis; IM, intestinal metaplasia; ns, not significant.

852 GENTA ET AL CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 7, No. 8 Figure 2. High power photomicrographs of H heilmannii on the surface gastric epithelium in one of the rare Helicobacter-infected patients in whom fundic gland polyps were found. Anti-Helicobacter immunohistochemical stain (magnification 60 ). Other Upper Gastrointestinal Pathology Patients with FGPs (Table 2) were twice as likely as those without to have a duodenal patch of gastric heterotopia (OR, 2.11; 95% CI, 1.58 2.81), but no more likely (OR, 1.36; 95% CI, 0.89 2.29) to have the same finding in the esophagus ( inlet patch ). The prevalence of duodenal intraepithelial lymphocytosis and villous atrophy consistent with celiac sprue was virtually identical in the 2 groups. Colonic Polyps and Neoplasms All patients with FGPs in the EGD/colonoscopy group were over 18 years of age. To avoid the bias resulting from the inclusion of children who had no FGPs and were unlikely to have colonic polyps, we only analyzed data from the 25,687 adults ( 18 years of age). Among these subjects, 1603 had FGPs; there were 1026 women (63.9%; median age 60 years) and 577 men (36.1%; median age 60 years). No FGPs were detected in 24,084 patients (median age 58 years) of whom 10,663 (44.3%) were men (median age 59 years) and 13,421 (65.7%) were women (median age 58 years). Table 3 depicts the prevalence of different types of colonic polyps (hyperplastic, serrated, and adenomatous), adenocarcinomas, and carcinoids in these 2 groups. In men with FGPs there was a slightly higher prevalence of hyperplastic polyps than in men without FGPs (38.7% vs 34.5%; OR, 1.20; 95% CI, 1.01 1.42, P.03); the rates of serrated polyps, adenomas, and neuroendocrine tumors were not significantly different between the 2 groups; however, there were only 2 colonic adenocarcinomas among the 577 men with FGPs ( 0.3%), compared with a prevalence of 1.4% in the 10,663 men with no FGPs (OR, 0.24; 95% CI, 0.06 0.97; P.05). Women with FGPs had a 42.3% prevalence of colonic adenomas, in contrast to 33.8% for those without FGPs (OR, 1.43; 95% CI, 1.26 1.63; P.001). The main indication for EGD was GERD in 237 (54.6%) of the 424 women with FGPs and adenomas, but only in 251 (44.0) of the 570 women without adenomas (OR, 1.53; 95% CI, 1.89 1.97; P.001). GERD was the indication in 136 (44.6%) of 305 men with FGPs and adenomas and in 135 (50.7%) of 266 men without (OR, 0.78; 95% CI, 0.56 1.09). Discussion The prevalence of FGPs has been reported to lie between 0.21% (in Italy) 19 and 2.0% (in Northern Ireland). 10 In the Table 3. Colonic Neoplasia in Patients With and Without Fundic Gland Polyps Type of colon polyp Men n 577; median age 60 Adults with FGP Adults without FGP N 1603; median age 59 N 24,084; median age 58 Women n 1026; median age 60 Men n 10,663; median age 59 Women n 13,421; median age 58 Hyperplastic polyps 225 (38.7); OR, 1.20; 95% CI, 321 (31.3) 3680 (34.5) 3940 (29.3) 1.01 1.42; P.03 Serrated polyps a 12 (2.1) 32 (3.1) 249 (2.3) 327 (2.4) Adenomas b 305 (52.6); OR, 1.15; 95% CI, 434 (42.3); OR, 1.43; 95% CI, 5226 (49.0) 4537 (33.8) 0.98 1.36; ns 1.26 1.63; P.0001 Adenocarcinoma c 2 (0.3); OR, 0.24; 95% CI, 11 (1.1) 152 (1.4) 141 (1.1) 0.06 0.97; P.05 Neuroendocrine tumors 0 (0.0) 0 (0) 12 (0.1) 21 (0.2) NOTE. Numbers and (percentages) of FGPs and non-fgp patients with different types of colonic polyps. Patients with no FGPs are arbitrarily assigned an OR of 1; only significant and borderline ORs are depicted. ns, not significant. a Includes all serrated polyps except hyperplastic polyps, serrated adenomas, and traditional serrated adenomas, with and without low- or high-grade dysplasia. b Includes all adenomas (tubular, tubulovillous, villous, traditional serrated adenomas, and mixed hyperplastic-tubular adenomas) with or without high-grade dysplasia. c Includes all adenocarcinomas detected in the colon, whether or not arising in a recognizable polyp.

August 2009 FUNDIC GLAND POLYPS, H. PYLORI, AND COLONIC NEOPLASIA 853 Irish study Dickey et al showed considerable foresight by commenting that it is too early to say that their prevalence will increase as H pylori incidence falls and proton pump inhibitor use grows, but improving endoscopic resolution will be sufficient in itself to bring their attention to the endoscopist more often. Twelve years later, in our study of almost 80,000 patients who underwent an EGD with gastric biopsies, 7.7% had 1 or more FGPs, which represented more than 70% of all excised polyps. The greater prevalence found in our population may be due to the widespread use of PPIs, which are used by a greater proportion of patients now than a decade or 2 ago, when the above studies were performed. It is also possible that endoscopists, and particularly American endoscopists, have become more aware of these lesions and have a greater interest in obtaining histopathologic confirmation of their benign nature. In an interesting, but somewhat unexplained comment, Dickey et al stated that [fundic gland polyps] reflect a histopathologist s viewpoint. 10 We interpret this to suggest that different criteria may be a reason for the variation in the reported prevalence rates. This is a legitimate concern that applies to any study where an interpretative finding (as opposed to an objective measurement) is involved. However, we are not aware of any data that would allow even to start speculating on the possible differences in diagnostic criteria between our group of pathologists and the pathologists involved with other studies. In our database, 154 of all patients with FGPs (2.5%) and 34 of the FGPs in the 1558 patients who had a concurrent colonoscopy (2.2%) had a diagnosis of early fundic gland polyp. This qualifier indicated that the glandular dilatations may not have been deemed to be sufficiently pronounced for a firm diagnosis, and the polyps may represent borderline lesions between true FGPs and pronounced PPI effect. Because the overall number of these possibly ambiguous lesions was small, they were equally distributed among the subgroups analyzed, and there are no clear diagnostic guidelines that would suggest their inclusion on a different category, we elected to include them in our analysis of FGPs. In other studies, FGPs have been reported to be 2 to 5 times more common in women than in men. 10,13,14,20 In our series (which consisted of 48,432 women and 30,477 men) 8.5% of the women and 6.4% of the men had FGPs, with a female to male ratio of 1.3 to 1, well below those previously reported. Although the fact that our diverse population from the continental United States and Puerto Rico was intrinsically different from those previously reported (which included uniform populations of Irish, Finnish, Scottish, Italian, and Japanese patients) cannot be discounted, it would appear more likely that the current widespread use of PPIs has contributed to an increase in FGPs in both men and women and has somehow almost equalized the prevalence. Further indirect evidence linking PPIs to the development of FGPs is the fact that GERD was the indication for EGD in almost half the patients who were found to have FGPs. Many patients with heartburn are using PPI therapy, and it is possible that those who undergo an endoscopy could be the ones who, being refractory to standard therapy, receive greater doses. A third of patients who did not have FGPs were reported to have an EGD for the investigation of GERD, but the difference between the 2 groups (OR,1.52; 95% CI, 1.45 1.61) is sufficiently large to warrant interest. The rarity of Helicobacter in the gastric mucosa of patients with FGP is well known to pathologists. Not only do our data support this impression and confirm the findings of other investigators, 10 13 but they also show a strong inverse association with chronic active gastritis, intestinal metaplasia, and, to a lesser extent, with chronic inactive gastritis. Interestingly, 2 of 6 FGPs with Helicobacter infection (30%) were infected with H heilmannii, whereas H heilmannii accounted for only 0.04% (39 of 9472) of patients who had Helicobacter gastritis in this series (OR, 120.93; 95% CI, 21.51 679.64). Thus, the typical stomach with an FGP is free of inflammation and has a relatively normal oxyntic mucosa, in which the only changes may be those associated with prolonged PPI use (mild focal or diffuse irregular dilatation of the oxyntic glands with parietal cells protruding into their lumen). 21 The antral mucosa in these patients may be normal, but with the widespread chronic use of nonsteroidal anti-inflammatory drugs, particularly in the older population, it is now common to see varying degrees of reactive gastropathy, a histological entity characterized by elongated and tortuous foveolae ( foveolar hyperplasia ), bundles of smooth muscle fibers projecting into the subepithelial zone, and regenerative changes of the surface epithelium. 22 In an Italian series, 5 of 70 patients with FGP (7.1%) also had gastric heterotopia in the duodenum. 14 While in our patients gastric heterotopia was rare (slightly less than 1%), it was found twice as often in patients with than in patients without FGPs. One possible explanation is that gastric heterotopia (a patch of oxyntic mucosa embedded in the duodenal mucosa) is also sensitive to PPIs, which cause glandular dilatation and could make the heterotopic patch more easily visible to endoscopists. The evaluation of the prevalence of polyps and neoplastic lesions in the colon of patients who have FGPs revealed some interesting associations: (1) men (not women) with FGPs had a slightly but significantly (P.03) greater prevalence of colonic hyperplastic polyps than those without FGPs; (2) women (not men) with FGPs had a significantly greater prevalence of colonic adenomas than those without FGPs (OR, 1.43; 95% CI, 1.26 1.63; P.0001); and (3) adenocarcinomas were discovered much less frequently in men with FGPs than in those without FGPs (OR, 0.24; 95% CI, 0.06 0.97; P.05). All other types of polyps were equally common in FGP patients and in those without FGPs. The weak correlation between colonic hyperplastic polyps and FGPs in men can be dismissed because of its low strength and the minimal clinical relevance of hyperplastic polyps. However, the higher prevalence of adenomas in women with FGPs remains difficult to interpret. The women with FGPs were 2 years older than those without FGPs (median age 60 years vs 58 years; P.001 by the Mann Whitney Rank Sum Test). In this population, regression analysis showed the prevalence of colonic adenomas to rise by 0.8% per year of age in subjects aged between 45 and 75 (data not shown). Thus, the prevalence in women with FGPs would be expected to be 1.6% greater than in those without FGPs, not 8.5% greater as we found. When we stratified women by 5-year age groups (data not shown), we found the increase to be distributed across all groups, but it was most pronounced (56.4% vs 46.1%) in women older than 60 years. In this study we had no access to specific information regarding PPI use in individual patients; it is, however, interesting to note that women with FGPs and adenomas were

854 GENTA ET AL CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 7, No. 8 significantly more likely to have an EGD because of GERD than women with FGPs but no adenomas (OR, 1.53; 95% CI, 1.89 1.97; P.001); no such difference was found in men or in patients of either sex without FGPs. While this may indirectly suggest a connection with PPI use, the possible relationship remains tenuous, also in light of studies showing that PPI use does not increase the risk of colorectal neoplasia. 23 It should be noted that our group of patients with FGPs was unusual with respect to its high percentage of bidirectional endoscopies: 25.7%, compared with approximately 10% reportedly found in general endoscopic practices. 24 In a retrospective study such as this, it is impossible to determine the precise clinical indications for this high rate of dual procedures, but this group, and perhaps the other groups in which an association of FGPs with colonic adenomas was reported, may have been either more or less representative of the usual population of patients undergoing endoscopy. Despite the possible true biological association of increased adenomas in women with FGPs, particularly in women over 60 years of age, we do not concur with the conclusion that a colonoscopy is indicated following the discovery of a sporadic FGP. Not only is the increase small, and we discovered no increased association with colonic adenocarcinoma, but patients in the most affected age group are already candidates for screening colonoscopy. References 1. Utsunomiya J, Maki T, Iwama T, et al. Gastric lesion of familial polyposis coli. Cancer 1974;34:745 754. 2. Watanabe H, Enjoji M, Yao T, et al. Gastric lesions in familial adenomatosis coli: their incidence and histologic analysis. Hum Pathol 1978;9:269 283. 3. Elster K, Eidt H, Ottenjann R, et al. The glandular cyst, a polypoid lesion of the gastric mucosa. Dtsch Med Wochenschr 1977;102: 183 187. 4. Declich P, Tavani E, Porcellati M, et al. Long-term omeprazole treatment and fundic gland polyps: a very authoritative proof against a link? Am J Gastroenterol 2001;96:1650. 5. Vieth M, Stolte M. Fundic gland polyps are not induced by proton pump inhibitor therapy. Am J Clin Pathol 2001;116:716 720. 6. el-zimaity HM, Jackson FW, Graham DY. Fundic gland polyps developing during omeprazole therapy. Am J Gastroenterol 1997; 92:1858 1860. 7. Choudhry U, Boyce HW Jr, Coppola D. Proton pump inhibitorassociated gastric polyps: a retrospective analysis of their frequency, and endoscopic, histologic, and ultrastructural characteristics. Am J Clin Pathol 1998;110:615 621. 8. Freeman HJ. Proton pump inhibitors and an emerging epidemic of gastric fundic gland polyposis. World J Gastroenterol 2008;14: 1318 1320. 9. Jalving M, Koornstra JJ, Wesseling J, et al. Increased risk of fundic gland polyps during long-term proton pump inhibitor therapy. Aliment Pharmacol Ther 2006;24:1341 1348. 10. Dickey W, Kenny BD, McConnell JB. Prevalence of fundic gland polyps in a western European population. J Clin Gastroenterol 1996;23:73 75. 11. Sakai N, Tatsuta M, Hirasawa R, et al. Low prevalence of Helicobacter pylori infection in patients with hamartomatous fundic polyps. Dig Dis Sci 1998;43:766 772. 12. Stachura J, Galazka K, Pirog K. Gastric fundic gland polyps (Elster s polyps) and Helicobacter pylori-induced gastritis. Pol J Pathol 1997;48:159 161. 13. Shand AG, Taylor AC, Banerjee M, et al. Gastric fundic gland polyps in south-east Scotland: absence of adenomatous polyposis coli gene mutations and a strikingly low prevalence of Helicobacter pylori infection. J Gastroenterol Hepatol 2002; 17:1161 1164. 14. Declich P, Tavani E, Ferrara A, et al. Sporadic fundic gland polyps: clinico-pathologic features and associated diseases. Pol J Pathol 2005;56:131 137. 15. Eidt S, Stolte M. Gastric glandular cysts investigations into their genesis and relationship to colorectal epithelial tumors. Z Gastroenterol 1989;27:212 217. 16. Jung A, Vieth M, Maier O, et al. Fundic gland polyps (Elster s cysts) of the gastric mucosa. A marker for colorectal epithelial neoplasia? Pathol Res Pract 2002;198:731 734. 17. Teichmann J, Weickert U, Riemann JF. Gastric fundic gland polyps and colonic polyps is there a link, really? Eur J Med Res 2008;13:192 195. 18. Lowry R. VassarStats. Available at: http://faculty.vassar.edu/ lowry/odds2x2.html. Accessed multiple times up to April 24, 2009. 19. Declich P, Isimbaldi G, Sironi M, et al. Sporadic fundic gland polyps: an immunohistochemical study of their antigenic profile. Pathol Res Pract 1996;192:808 815. 20. Marcial MA, Villafana M, Hernandez-Denton J, et al. Fundic gland polyps: prevalence and clinicopathologic features. Am J Gastroenterol 1993;88:1711 1713. 21. Stolte M, Bethke B, Ruhl G, et al. Omeprazole-induced pseudohypertrophy of gastric parietal cells. Z Gastroenterol 1992;30: 134 138. 22. Genta RM. Differential diagnosis of reactive gastropathy. Semin Diagn Pathol 2005;22:273 283. 23. Robertson DJ, Larsson H, Friis S, et al. Proton pump inhibitor use and risk of colorectal cancer: a population-based, case-control study. Gastroenterology 2007;133:755 760. 24. Urquhart J, Eisen G, Faigel DO, Mattek N, Holub J, Lieberman DA. A closer look at same-day bidirectional endoscopy. Gastrointest Endosc 2009;69:271 277. Reprint requests Address requests for reprints to: Robert M. Genta, MD, Dallas VAMC, Pathology and Medicine (Gastroenterology), 4500 South Lancaster Road, Dallas, Texas 75216. e-mail: robert.genta@utsouthwestern.edu; fax: (214) 596-2297. Conflicts of interest The authors disclose no conflicts.