Gastric Cardiac Polyps a Clinicopathologic Study of 330 Cases

ORIGINAL ARTICLE

Gastric Cardiac Polyps A Clinicopathologic Study of 330 Cases

Shelby D. Melton, MD*w and Robert M. Genta, MD*w

uring the last few years, endoscopists have become Abstract: As endoscopists have become more skilled in sampling Dincreasingly skilled in correctly sampling the gastro- the gastroesophageal junction, pathologists are being increasingly esophageal junction, which they do in approximately 1 of challenged to characterize previously unknown or neglected every 3 esophagogastroduodenoscopies.22 The availabil- findings. One such example is the cardiac polyp. Originally ity of increasing numbers of squamocolumnar biopsy described in the radiology literature as the sentinel fold, the first specimens has provided pathologists with new challenges histologic descriptions of polyps at the gastroesophageal junction and also the opportunity to characterize previously did not appear until less than a decade ago. Current clinico- unknown or neglected findings. One example is the pathologic information is limited and somewhat conflicting. This recognition of small foci of pancreatic tissue found in study was designed to define the clinical, endoscopic, and the mucosa of the gastroesophageal junction in 5% to histopathologic associations in patients with cardiac polyps. Using 16% of biopsies.6,10,21 Another opportunity has emerged an electronic database, we extracted information on all patients with the rediscovery of inflammatory polypoid lesions at who had a distal esophageal or esophagogastric junctional biopsy the esophagogastric junction, known to some radiologists during a 24-month period. We then reviewed the slides of 330 and endoscopists as the “sentinel fold” (Fig. 1A). adult patients diagnosed with a cardiac polyp and used Originally reported by radiologists as exuberant growths semiquantitative or qualitative scales to score foveolar hyperpla- in the pouch of patients with hiatal hernia,3,9 they were sia, inflammation, erosion or ulcers, epithelial type, and metapla- thought to be related to gastroesophageal reflux disease sia. As controls we used 120,487 patients who had biopsies from (GERD), to the point of being named “the reflux the same anatomic sites during the same period, but were not gastroesophageal polyp.”2,11,13,16,23 Since then, there have diagnosed with a cardiac polyp. There were no significant been reports of individual cases or small series, some of differences among any clinical indications for esophagogastro- which refuted the link with reflux.18 The first systematic duodenoscopy between study and control patients. Endoscopi- histopathologic analyses of these lesions did not come up cally, a polyp or nodule at the gastroesophageal junction was until the early 2000s.1,20 noted in 59.1% of the patients who had a histopathologic These studies, conducted in dissimilar populations, diagnosis of cardiac polyp. Histologically, Barrett mucosa, active reported a variety of clinicopathologic associations, did esophagitis, and Helicobactor pylori gastritis were all significantly not always include a control group, and were not always less common in patients with a cardiac polyp than in controls. in agreement with one another. During the last few years, Although relatively infrequent, synchronous hyperplastic polyps we have collected data on such lesions, which we refer to elsewhere in the stomach were significantly more common in as “cardiac polyps,” from a large U.S. population. To patients than in controls. In conclusion, this large series suggests elucidate the histopathologic, clinical, and endoscopic that cardiac polyps are rare but histologically distinct lesions. associations of cardiac polyps, we compared the findings They are benign and are not uniquely associated with esophagitis, in 330 patients with a cardiac polyp to those of 120,000 Barrett esophagus, gastroesophageal reflux disease, reactive individuals who had biopsies from the gastroesophageal gastropathy, or gastritis, with or without H. pylori. junction without evidence of such lesions. Key Words: cardia, hyperplastic polyp, gastroesophageal junction, foveolar hyperplasia, sentinel fold MATERIALS AND METHODS (Am J Surg Pathol 2010;34:1792–1797) Study Setting This study was conducted at Caris Life Sciences, a From the *Caris Research Institute, Caris Life Sciences, Irving; and laboratory that receives specimens from gastroenterolo- wDepartment of Pathology, Veterans Affairs North Texas Health Care System, University of Texas Southwestern Medical Center, gists who operate in private outpatient endoscopy centers Dallas, TX. across the United States. Biopsies are interpreted by Conflict of interest disclosure: Robert Genta is an employee of Caris Life a group of gastrointestinal pathologists who have Sciences. Shelby Melton has no financial interest to declare. achieved a high level of diagnostic uniformity through a Correspondence: Robert M. Genta, MD, Division of Gastrointestinal Pathology, Caris Life Sciences, 6655 North MacArthur Blvd, Irving, predetermined approach to specimen handling, diagnostic TX 75039 (e-mail: [email protected]). criteria, and terminology. Consensus is maintained Copyright r 2010 by Lippincott Williams & Wilkins and updated through daily multiheaded microscope

1792 | www.ajsp.com Am J Surg Pathol Volume 34, Number 12, December 2010 Am J Surg Pathol Volume 34, Number 12, December 2010 Gastric Cardiac Polyps

FIGURE 1. A, Endoscopic appearance of a polyp at the gastroesophageal junction. No erosion or ulceration is seen on the surface of the polyp, and there are no features suggestive of a malignant process. The surrounding mucosa is unremarkable. Reproduced with permission from Dr Timothy Huggins, Weatherford, TX. B, The cardia location is confirmed by the presence of squamous mucosa (arrow). The gastric pits are elongated and irregular, and there is a brisk inflammatory component within the lamina propria. No dilated oxyntic glands are seen (hematoxylin and eosin, original magnification 40 Â). C, The dilated foveolae confer a villiform appearance to the columnar mucosa. The foveolar cells have benign cytologic features, and the squamous mucosa (S) is unremarkable. There is a mild chronic inflammatory infiltrate within the lamina propria without active inflammation (hematoxylin and eosin, original magnification 200 Â). D, This large cardiac polyp shows the characteristic marked foveolar hyperplasia (F), and focal erosion of the surface (arrow). The lamina propria is edematous (E), with little chronic and no active inflammation (hematoxylin and eosin, original magnification 40 Â). conferences, monthly didactic and journal review con- The study was approved by the Caris Life Sciences ferences, a terminology review committee, and ongoing Institutional Review Board. As all the data in this study quality assurance reviews. were collected entirely by reviewing existing deidentified r 2010 Lippincott Williams & Wilkins www.ajsp.com | 1793 Melton and Genta Am J Surg Pathol Volume 34, Number 12, December 2010 records, no direct contact with either patients or providers TABLE 1. Demographic Data, Indications for Endoscopy, and was involved, and no individual information is revealed in Endoscopic Impressions in Patients With and Without a any form, this study was considered exempt from the need Cardiac Polyp for informed consent from participants. Patients With Patients Without Cardiac Polyp Cardiac Polyp Sources of Data (n = 330) (n = 120,487) OR (95% CI) Data were extracted from Caris electronic database, which include demographic and clinical information Age (y) 60 (17-94) 57 (17-102) Women 168 (50.9) 63,494 (52.7) NS for each patient, a summary of the endoscopic findings or Indication for endoscopy the entire endoscopic report, the site of origin of each Dyspepsia 35 (10.6) 13,731 (11.4) NS specimen, and the histopathologic report for each biopsy. Dysphagia 64 (19.4) 23,624 (19.6) NS To extract eligible cases, we used simple verbal queries Reflux 90 (27.3) 36,372 (30.2) NS Noncardiac 12 (3.6) 4508 (3.7) NS and Boolean logic to electronically review all pathology chest pain reports from patients who had at least one biopsy from Endoscopic impressions the distal esophagus, gastroesophageal junction, or Polyp or 195 (59.1) 5112 (4.2) 13.9 (12.7-15.3) gastric cardia submitted to the Caris Life Sciences nodule (P<0.0001) between January 1, 2008 and December 31, 2009. Barrett 116 (35.2) 58,813 (48.8) 0.57 (0.45-0.71) esophagus (P<0.0001) Esophagitis 104 (31.5) 39,481 (32.8) NS Clinical Indications and Endoscopic Impressions Hiatal 101 (30.6) 33,867 (28.1) NS Esophageal indications for endoscopy considered hernia pertinent to the study included GERD or reflux, dysphagia Only significant OR values are reported. or odynophagia, dyspepsia, and noncardiac chest pain. CI indicates confidence interval; NS, not significant; OR, odds ratio. Relevant endoscopic findings in the esophagus were (1) nodule or polyp at the gastroesophageal junction or distal esophagus; (2) esophagitis; (3) Barrett esophagus (including “rule out,” “suspicious for,” or “consistent with” Barrett, and without a cardiac polyp. There were no significant and “salmon-colored mucosa”); and (4) hiatal hernia. differences between patients and controls in age, sex, or clinical indications for esophagogastroduodenoscopy. Histopathologic Evaluation The authors retrieved and reviewed the slides after Endoscopic Impressions of the Esophagus identifying the study cases. For each case we used a A nodule or polyp in the cardia, distal esophagus, or semiquantitative scale from 0 (absent) through 3 (marked at the gastroesophageal junction was reported endosco- or “severe”) modeled on the updated Sydney System pically in 195 of the 330 patients with a histologic cardiac scales5 to score the following features: (1) foveolar polyp (59.1%). Figure 1A shows an endoscopic view of hyperplasia, (2) chronic inflammation, and (3) active a polyp at the gastroesophageal junction. Although there inflammation. In addition, the following features were was variation in the size of the polyps, they were rarely noted as present or absent: (1) lymphoid aggregates, described endoscopically as measuring 1 cm or more. An (2) intestinal metaplasia, (3) dysplasia, and (4) neoplasia. impression of Barrett esophagus (including the descrip- We also noted whether the polyp consisted of mucous tion of “salmon-colored mucosa”) was recorded in fewer epithelium, oxyntic epithelium, or a combination of the patients with a cardiac polyp than in those without two, and whether squamous epithelium was present in (34.8% vs. 48.8%, respectively; P<0.001). Esophagitis the biopsy specimen. Immunohistochemical staining for was seen at comparable frequencies in patients with and Helicobactor pylori was carried out on any accompanying without a cardiac polyp (31.5% vs. 32.7%). In addition, a gastric biopsies (Cell Marque, Rocklin, CA). Gastric hiatal hernia was noted in a similar percentage of patients polyps were diagnosed according to previously published, with and without a cardiac polyp (30.6% vs. 28.1%), as widely accepted criteria.4 was a mention of an irregular Z line (data not shown).

RESULTS Histologic Findings in Cardiac Polyps Figures 1B to D show the salient histologic findings Demographics and Clinical Data of cardiac polyps. The defining feature of these lesions There were 120,817 adult patients with distal or was prominent (polypoid) foveolar hyperplasia, charac- junctional esophageal biopsies during the study period terized by elongated and tortuous foveolae and dilatation (median age, 57 y; range, 17 to 102; 52.7% women). of mucous glands. Mitoses were generally absent. The A cardiac polyp was diagnosed in 330 patients (median stroma was frequently edematous; however, prominent age, 60 y; range, 17 to 94; 50.9% women). The control smooth muscle bundles, collagen deposition, or fibrosis in group (patients with no diagnosis of cardiac polyp) the lamina propria were not findings associated with these included 120,487 patients (median age, 57 y; range, 17 to lesions. Table 2 summarizes the histologic scores recorded 102; 52.7% women). Table 1 lists the demographic during review of the study cases. Erosions were identified information, clinical indications for endoscopy, and on the surface epithelium in half of the polyps; however, endoscopic impressions of the esophagus in patients with polypoid fragments of granulation tissue were not

1794 | www.ajsp.com r 2010 Lippincott Williams & Wilkins Am J Surg Pathol Volume 34, Number 12, December 2010 Gastric Cardiac Polyps

prevalence of histologically confirmed Barrett esophagus TABLE 2. Summary of Histologic Characteristics and Corresponding Scores in Cardiac Polyps in 330 Patients was almost 8 times lower in patients with a cardiac polyp (1.8%) than in those without (13.4%). The prevalence of Patients With Cardiac Polyp (n = 330) % eosinophilic and candida esophagitis was not significantly different in patients and controls. There were no Endoscopically noted polyp or nodule 195 59.1 cytomegalovirus-related or Herpes simplex virus-related Squamous mucosa identified 189 57.3 Foveolar hyperplasia diagnoses in patients with a cardiac polyp, compared with Mild 103 31.2 20 viral infections (16 cytomegalovirus-related and 4 Moderate 137 41.5 Herpes simplex virus-related diagnoses) seen in control Marked 89 27.0 patients. Neither invasive carcinomas nor Barrett epithe- Chronic inflammatory infiltrate Mild 119 36.1 lium with any grade of dysplasia were identified in any Moderate 144 43.6 patients with a cardiac polyp. Marked 67 20.3 Active inflammation of the epithelium Histologic Findings in Gastric Biopsies None 135 40.9 Mild 99 30.0 A total of 241 patients with a cardiac polyp and Moderate 56 17.0 76,378 control patients also had at least 1 simultaneous Marked 39 11.8 biopsy from the stomach (excluding the cardia). Relevant Pure mucinous epithelium 263 79.7 findings (depicted in Table 4) included H. pylori infection, Oxyntic epithelium component 67 20.3 Lymphoid aggregates in lamina propria 38 11.5 which was significantly less common in patients with Foci of pancreatic metaplasia 17 5.2 cardiac polyps than in those without, and the presence of Mucosal erosion/ulceration 162 49.1 synchronous hyperplastic polyps elsewhere in the sto- Intestinal metaplasia 36 10.9 mach, 3 times more common in patients with cardiac Multilayered epithelium 16 4.8 polyps than in controls. There were no significant differences between the 2 groups with regard to other gastric histopathologic diagnoses. included in the analysis. Intestinal metaplasia was detected in 36 of the cardiac polyps (10.9%), twice the rate of focal intestinal metaplasia at the gastroesophageal DISCUSSION junction found in controls (5.3%), but similar to that Initial reports of polyps at the gastroesophageal reported in the literature.14,17,19 One polyp had cytologic junction consisted of either individual cases or small series features that were interpreted at a consensus conference published in the radiology and gastroenterology litera- 3,8,9,13,16,18 as representing reactive atypia; no cases of dysplasia or ture. The first systematic histopathologic char- 1 carcinoma were found in patients with cardiac polyps, acterization of these lesions was done by Abraham et al either within the polyp or in the gastroesophageal in 2001. The investigators studied clinical, endoscopic, junctional region. and histologic data from 30 hyperplastic polyps from the esophagus or esophagogastric junction in 27 patients. The second major series studied 31 individuals from Fin- Histologic Findings in Esophageal Biopsies land.20 Recently, a series of 46 cases of hyperplastic A summary of the histologic findings in additional polyps at the gastroesophageal junction was reported in esophageal biopsies from both groups of patients is found abstract form.12 Differences in the patient populations in Table 3. Mucosal erosion or ulceration was similar in studied in these series (children vs. adults, inclusion or patients with cardiac polyp and controls. However, active exclusion criteria, comparison with controls, correlation esophagitis with characteristic changes of reflux was with biopsies from elsewhere in the upper gastrointestinal diagnosed significantly less frequently in patients with a tract) and differences in histopathologic criteria, have cardiac polyp (14.5%) than in those without (35.5%). The yielded a varied spectrum of clinical and pathologic

TABLE 3. Histopathology in the Esophagus in Patients With and Without a Cardiac Polyp Patients With Cardiac Patients Without Cardiac Polyp (n = 330) Polyp (n = 120,487) OR (95% CI) Histopathologic ﬁndings in the esophagus Active esophagitis 48 (14.5) 42,713 (35.5) 0.31 (0.23-0.42) (P<0.0001) Barrett mucosa 6 (1.8) 16,153 (13.4) 0.12 (0.05-0.27) (P<0.0001) Erosion/ulceration 12 (3.6) 3717 (3.1) NS Eosinophilic esophagitis 3 (0.9) 2813 (2.3) NS Candida esophagitis 4 (1.2) 1241 (1.0) NS

Only significant OR values are reported. CI indicates confidence interval; NS, not significant; OR, odds ratio.

r 2010 Lippincott Williams & Wilkins www.ajsp.com | 1795 Melton and Genta Am J Surg Pathol Volume 34, Number 12, December 2010

TABLE 4. Histopathologic Findings in Gastric Biopsies From Patients With and Without a Cardiac Polyp Patients With Cardiac Polyp Patients Without Cardiac Polyp (n = 241) (n = 76,378) OR (95% CI) Histopathologic ﬁndings in the stomach Gastric hyperplastic polyp 10 (4.1) 1007 (1.3) 3.24 (1.71-6.12)* Fundic gland polyp 12 (5.0) 7035 (9.2) 0.52 (0.29-0.92) (P = 0.03) Helicobacter pylori gastritis 7 (2.9) 5601 (7.3) 0.38 (0.18-0.80) (P = 0.012) Reactive gastropathy 32 (13.3) 12,796 (16.8) NS Chronic inactive gastritis 9 (3.7) 4244 (5.6) NS Helicobactor pylori-negative chronic 3 (1.2) 1187 (1.6) NS active gastritis

Only significant OR values are reported. *Sample too large for the Fisher exact test. CI indicates confidence interval; NS, not significant; OR, odds ratio.

associations for these junctional polyps. This study polyp (16.7%). These findings are in agreement with both extracted data on the participants from a very large the series of Voutilainen12 and Long.20 They also provide population of community-based private gastroenterology additional evidence that the foveolar hyperplasia occur- practices in 42 states. It also included a control group of ring at the gastroesophageal junction is unlikely to have a more than 120,000 individuals who underwent similar chemical reactive pathogenesis. procedures. All biopsies were obtained and read during The defining characteristics of this series of lesions a 2-year period, leaving less space for the bias associated were foveolar hyperplasia and a chronic inflammatory with changing endoscopic practices. The well-defined set infiltrate. Erosions and ulcers were found in half of the of histopathologic criteria that were used also provided polyps and some degree of active inflammation in about an additional element of uniformity to our data. two-thirds of them. The polypoid appearance was likely Barrett esophagus was suspected endoscopically in a caused by the foveolar hyperplasia, the edema, the third of the patients with a cardiac polyp, compared with inflammation, and less commonly by additional compo- half of the controls. These endoscopic impressions are in nents, such as prominent lymphoid aggregates (in 11.5% agreement with our histopathologic findings: Barrett of the polyps) and foci of pancreatic metaplasia (in mucosa was confirmed histologically in only 6 (1.8%) 5.1%).21 Simply calling these lesions foveolar hyperplasia patients with a cardiac polyp, and in 13.4% of controls. (as in study of Voutilainen)20 is, in our view, inadequate These data contrast with a recent report in abstract form, to convey the topographic information that the diagnosis which detected a significant association with Barrett should provide. Furthermore, the ingrained association of esophagus (15 of 46; 33%).12 Furthermore, as no definite foveolar hyperplasia with reactive gastropathy may risk dysplasia (low or high grade) was identified in any of the giving the erroneous impression that chemical injury may patients, evidence would suggest that cardiac polyps are be involved in their pathogenesis. Therefore, we suggest not preneoplastic lesions. However, further studies to that the terms “cardiac polyp” or “gastroesophageal investigate clonality and mutation status and follow-up junctional polyp” be used to refer to these lesions. data, may help determine whether there is a risk of In conclusion, this large series suggests that cardiac adverse clinical behavior. polyps are rare but histologically distinct lesions with a The presence of a cardiac polyp was inversely frequent endoscopic correlate. They are benign and show associated with H. pylori gastritis (odds ratio, 0.38; 95% no significant association with esophagitis, Barrett confidence interval, 0.18-0.81), and there was no connec- esophagus, GERD, reactive gastropathy, or gastritis, tion with chronic gastritis. Patients with a cardiac polyp with or without H. pylori. were 3 times more likely to have a gastric hyperplastic polyp elsewhere in the stomach than controls. However, ACKNOWLEDGMENTS the rarity of this finding (occurring in only 10 of the 330 The authors thank Dr Timothy Huggins, the Surgery patients) makes it difficult to invoke a common patho- Center at Park City, Weatherford, Texas, for providing the genesis linking them together. Gastric foveolar hyperpla- endoscopic image of a cardiac polyp (Fig. 1A). They also sia is one of the hallmarks of reactive (or chemical) acknowledge the competence and support of the technical gastropathy, a set of histopathologic findings observed in and administrative staff at Caris Life Sciences. a portion of patients with bile reflux and those who use nonsteroidal anti-inflammatory drugs.7,15 As a complete medication history was not available to us, we used REFERENCES reactive gastropathy as a surrogate marker for nonster- 1. Abraham SC, Singh VK, Yardley JH, et al. Hyperplastic polyps of the esophagus and esophagogastric junction: histologic and clin- oidal anti-inflammatory drug use. We found an almost icopathologic findings. Am J Surg Pathol. 2001;25:1180–1187. identical prevalence of reactive gastropathy in patients 2. Bach KK, Postma GN, Koufman JA. Esophagitis with an with a cardiac polyp (13.2%) and those with no cardiac inflammatory polyp. Ear Nose Throat J. 2002;81:824.

1796 | www.ajsp.com r 2010 Lippincott Williams & Wilkins Am J Surg Pathol Volume 34, Number 12, December 2010 Gastric Cardiac Polyps

3. Bleshman MH, Banner MP, Johnson RC, et al. The inflam- 13. Mohandas KM, Santhi SV, Krishnamurthy S, et al. Sentinel polyp and matory esophagogastric polyp and fold. Radiology. 1978;128: fold: resolution following drug therapy. Endoscopy. 1992;24:293–294. 589–593. 14. Morales TG, Sampliner RE, Bhattacharyya A. Intestinal metaplasia 4. Carmack SW, Genta RM, Saboorian MH. The current spectrum of of the gastric cardia. Am J Gastroenterol. 1997;92:414–418. gastric polyps: a one-year national study of over 120,000 patients. 15. Pusztaszeri MP, Genta RM, Cryer BL. Drug-induced injury in the Am J Gastroenterol. 2009;104:1524–1532. gastrointestinal tract: clinical and pathologic considerations. Nat 5. Dixon MF, Genta RM, Yardley JH, et al. Classification and grading Clin Pract Gastroenterol Hepatol. 2007;4:442–453. of gastritis. The updated Sydney System: International Workshop 16. Rabin MS, Bremner CG, Botha JR. The reflux gastroesophageal on the Histopathology of Gastritis, Houston 1994. Am J Surg polyp. Am J Gastroenterol. 1980;73:451–452. Pathol. 1996;20:1161–1181. 17. Spechler SJ. Intestinal metaplasia at the gastroesophageal junction. 6. El-Serag HB, Graham DY, Rabeneck L, et al. Prevalence and Gastroenterology. 2004;126:567–575. determinants of histological abnormalities of the gastric cardia in 18. Van der Veer LD, Kramer K, Relkin R, et al. The esophagogastric volunteers. Scand J Gastroenterol. 2007;42:1158–1166. polyp-fold complex. Am J Gastroenterol. 1984;79:918–920. 7. Genta RM. Differential diagnosis of reactive gastropathy. Semin 19. Voutilainen M, Farkkila M, Juhola M, et al. Specialized columnar Diagn Pathol. 2005;22:273–283. epithelium of the esophagogastric junction: prevalence and associa- 8. Glassman M, Bostwick HE, Godine L, et al. Endoscopic removal of tions: the Central Finland Endoscopy Study Group. Am J inflammatory esophagogastric polyps in children. J Pediatr Gastro- Gastroenterol. 1999;94:913–918. enterol Nutr. 1991;13:110–114. 20. Voutilainen M, Juhola M, Farkkila M, et al. Foveolar hyperplasia 9. Janson R. Polyps in the prolapsed gastric part of an esophagogastric at the gastric cardia: prevalence and associations. J Clin Pathol. hiatal hernia. Fortschr Geb Rontgenstr Nuklearmed. 1971;114: 2002;55:352–354. 713–715. 21. Wang HH, Zeroogian JM, Spechler SJ, et al. Prevalence and 10. Johansson J, Hakansson HO, Mellblom L, et al. Pancreatic acinar significance of pancreatic acinar metaplasia at the gastroesophageal metaplasia in the distal oesophagus and the gastric cardia: junction. Am J Surg Pathol. 1996;20:1507–1510. prevalence, predictors and relation to GORD. J Gastroenterol. 22. Wang KK, Sampliner RE. Updated guidelines 2008 for the 2010;45:291–299. diagnosis, surveillance and therapy of Barrett’s esophagus. Am 11. Kato S, Ozawa A, Shibuya H, et al. Inflammatory esophagogastric J Gastroenterol. 2008;103:788–797. polyp and fold in an adolescent. Acta Paediatr Jpn. 1993;35:53–56. 23. Zitsman JL, Schullinger JN, Berdon WE. Inflammatory esophago- 12. Long KB, Odze RD. Cardiac polyps: a clinicopathologic study of 49 gastric polyps: resolution following antireflux surgery. J Pediatr patients. Mod Pathol. 2010;23(suppl 1):154A. Surg. 1988;23:1016–1017.

r 2010 Lippincott Williams & Wilkins www.ajsp.com | 1797