The term juvenile polyp was used in 1957 by Horrilleno. Juvenile Polyps: Recurrence in Patients With Multiple and Solitary Polyps

Transcription

1 CLINICAL GASTROENTEROLOGY AND HEPATOLOGY 2010;8: Juvenile Polyps: Recurrence in Patients With Multiple and Solitary Polyps VICTOR L. FOX,* STEPHEN PERROS,* HONGYU JIANG,*, and JEFFREY D. GOLDSMITH, *Division of Gastroenterology and Nutrition, Clinical Research Program, Department of Pathology, Children s Hospital Boston, Boston, Massachusetts; and Department of Pathology, Beth Israel Deaconess Medical Center, Boston, Massachusetts BACKGROUND & AIMS: Juvenile polyps are benign hamartomas with neoplastic potential that are the most frequent gastrointestinal polyp of childhood. Most information about juvenile polyps in childhood comes from small published series that lack detailed outcome data. We sought to identify a large cohort of children with one or more polyps and analyze clinical characteristics, including polyp recurrence, which might contribute to the development of management guidelines. METHODS: A retrospective chart review study of patients with juvenile polyps of the colon was performed. Cases were identified by searching a single hospital pathology database from 1990 to 2009 for the diagnosis of juvenile polyps. Recorded information included basic demographics, family history, genetic testing, and colonoscopy and pathology reports. RESULTS: A total of 257 children (median age, 5.6 y; 61.5% male) with juvenile polyps were identified. Among 192 patients who underwent complete colonoscopy at initial diagnosis, 117 (60.9%) had a single polyp, 75 (39.1%) had multiple polyps, 8 (4.2%) had polyps restricted to the right colon, and a total of 1653 polyps were found during 350 colonoscopy examinations. Polyps recurred in 21 of 47 (44.7%) patients after initial eradication, including 3 (16.7%) of 18 presenting with a single polyp. Neoplasia was found in 10 of 257 (3.9%) patients (right colon in 7 patients). Germline DNA abnormalities in mothers against decapentaplegic Drosophila (SMAD4), bone morphogenetic protein receptor 1A (BMPR1A), and phosphatase and tensin homolog (PTEN) were detected in 10 of 23 (43.5%) patients with multiple polyps. CONCLUSIONS: Recurrent polyp formation is common in children with juvenile polyps and occurs in patients with multiple and solitary polyps. Standardized protocols for detecting polyp recurrence, associated gene mutations, and neoplasia should be developed for children with juvenile polyps. Keywords: Juvenile Polyp; Juvenile Polyposis Syndrome; Neoplasia; Dysplasia; Hamartomas; Hereditary Hemorrhagic Telangiectasia; Pediatric; Gastrointestinal Bleeding. View this article s video abstract at The term juvenile polyp was used in 1957 by Horrilleno et al 1 to describe a distinctive nonadenomatous polyp commonly found in the rectum and colon of young children that contains dilated cystic spaces, exuberant lamina propria inflammation including eosinophils, marked vascularity, and areas of ulceration. Juvenile polyps are the predominant gastrointestinal polyp of childhood, exceeding 90% in some series. 2,3 The reported incidence of colonic polyps in children has ranged from 0.08% to 3.7%, with higher rates reported in symptomatic children 1,4 6 and with a rate of 1.1% reported in an autopsy series of 449 children. 7 Juvenile polyps are classified as benign hamartomatous lesions with neoplastic potential The presence of multiple juvenile polyps may indicate a premalignant condition commonly named juvenile polyposis coli or juvenile polyposis syndrome (JPS) In contrast, single or solitary juvenile polyps generally are considered benign sporadic lesions that confer little to no future risk of malignancy. 15 In practice, however, these 2 entities are not always easily distinguishable, resulting in confusion about proper management. A diagnosis of JPS is firmly established by the presence of one or more juvenile polyps associated with a family history of juvenile polyps or documented germline mutations in mothers against decapentaplegic Drosophila (SMAD4) or bone morphogenetic protein receptor 1a (BMPR1A). 16,17 Germline mutations in phosphatase and tensin homolog (PTEN) also are found in patients with juvenile polyps, although some investigators prefer to classify these patients separately because of their more complex phenotype. 18 Proposed clinical working definitions also have included the presence of 3 or more juvenile polyps 13 or more than 5 juvenile polyps 12 with or without other identified risk factors. These working definitions are derived from the observation of increased association in such patients with dysplasia and colorectal adenocarcinoma. According to these criteria, the presence of 1 or 2 polyps represents a sporadic low-risk condition not warranting future surveillance. The patient s age at the time of presentation and the time interval over which multiple polyps may be discovered have not been factored into any working clinical definitions. Most information about the risk of malignancy associated with JPS is derived from adult patients. The cumulative lifetime risk of colorectal cancer in patients with JPS has been estimated to range from 38.7% to 68%, 12,19 with a mean age at diagnosis in the fourth to fifth decade of life. Severe dysplasia or adenocarcinoma in situ has been reported in two 12-year-old children with juvenile polyps 20,21 and advanced adenocarcinoma has been found in children as young as 15 years of age. 10,12 Malignancy rarely has been reported in solitary juvenile polyps. 10 The data concerning juvenile polyps in childhood comes from small published series that frequently include non juvenile-type polyps. In addition, these studies used different criteria for the diagnosis of JPS and offer little or no detail about polyp recurrence. The primary aim of our study was to identify Abbreviations used in this paper: BMPR1A, bone morphogenetic protein receptor 1A; JPS, juvenile polyposis syndrome; PTEN, phosphatase and tensin homolog; SMAD4, mothers against decapentaplegic Drosophila by the AGA Institute /$36.00 doi: /j.cgh

2 796 FOX ET AL CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 8, No. 9 a large cohort of children with any number of juvenile polyps for detailed clinical characterization. We particularly sought to investigate patterns of polyp number, location, and neoplasia, and polyp recurrence over time in patients who underwent total colonoscopy including complete eradication of polyps during serial examinations. Patients and Methods Patients with juvenile polyps were identified by searching the Children s Hospital Boston Pathology Database for all reports that included the term juvenile polyp in the final diagnosis for specimens submitted from January 1990 to May Hospital records were reviewed retrospectively for each patient after approval by the hospital committee on clinical investigation. Recorded information included basic demographics, family history, genetic testing, and details from endoscopy and pathology reports noting polyp number, location, and histologic description. Patients with a background diagnosis of chronic inflammatory bowel disease were excluded. Polyp number and location were analyzed for patients who underwent complete colonoscopy at the time of initial diagnosis. Polyps that clearly were recognized endoscopically and confirmed by histology were counted. Equivocal endoscopic or histologic findings were excluded. A colonoscopy was considered to be complete if specifically stated by the endoscopist or if it included examination of the cecum or terminal ileum. The right and left colon were defined as segments proximal or distal to the splenic flexure, respectively. A diagnosis of neoplasia required the presence of dysplasia (low or high grade) or invasive adenocarcinoma. The patient age at diagnosis was determined by the date of the first diagnostic specimen. Polyp recurrence was measured in patients who had complete polyp removal or eradication during each successive complete colonoscopy. Statistical analysis of data was performed using the SAS 9.2 software program (SAS Institute, Inc, Cary, NC). Frequency and proportions were used to summarize categoric characteristics. Median values and interquartile range were used to summarize quantitative characteristics. Categoric outcomes were compared between groups using the chi-square test or the Fisher exact test where appropriate. The Wilcoxon rank sum test was used to compare quantitative outcomes between groups. A P value of less than.05 was considered statistically significant. Results Polyp Number and Location A total of 257 children were identified with one or more juvenile polyps. The median age at diagnosis was 5.6 years (interquartile range, y) and 158 (61.5%) were boys. A Table 1. Patient Characteristics Total children 192 Total polyps 1653 Male sex, % 64.1 Median age at diagnosis, y (interquartile range) 5.8 ( ) Location of polyps, n (%) Left colon only 128 (66.6) Left and right colon 56 (29.2) Right colon only 8 (4.2) Table 2. Number of Polyps Found During Initial Colonoscopy No. polyps No. patients (% total) (60.9) 2 15 (7.8) 3 21 (10.9) 4 10 (5.2) 5 7 (3.6) 6 4 (2.1) 7 2 (1.0) 9 3 (1.6) 10 1 (0.5) 11 1 (0.5) 12 3 (1.6) 14 1 (0.5) 15 2 (1.0) 17 5 (2.6) total of 192 (74.7%) patients from this original cohort underwent complete colonoscopy at the time of initial diagnosis and a total of 350 completed colonoscopy procedures. This subcohort was subjected to more detailed retrospective analysis. The age and gender characteristics for this subcohort were comparable with the initial total group. Tables 1 and 2 show patient characteristics, polyp location, and number of polyps found at the time of initial diagnosis. In this subgroup, 117 (60.9%) had a single polyp and 75 (39.1%) had more than 1 polyp. Although polyps were found predominantly in the left colon, 8 (4.2%) patients had polyps limited to the right colon and 7 of these had a single polyp. Sixty (31.3%) patients had 3 polyps or more and 22 (11.4%) patients had more than 5 polyps. Patients with more than one polyp were more likely than those with a single polyp to have a family history of a first- or second-degree relative with polyps (any type) or colon cancer (P.006). No significant difference was found between these 2 patient groups with respect to sex, age, or race. Polyp Recurrence Repeat surveillance colonoscopy was performed in 62 of 192 (32.3%) patients found to have one or more polyps during the initial complete colonoscopy. This included 18 (15.4%) of the 117 patients with a single polyp and 44 (58.7%) of the 75 patients with multiple polyps. Figure 1 shows the characteristics of the cohort of patients who underwent sequential follow-up complete colonoscopy examinations. Most patients had polyps completely eradicated or removed during each examination. Median time intervals between examinations ranged from 12.4 to 19.8 months. The majority of patients returned for each of the repeat examinations because of recommendations by the primary gastroenterologist based on prior findings and suspected continued polyp formation rather than recurrent symptoms such as bleeding. Recurrent bleeding was included as an indication in 26 of 62 patients during the first follow-up examination, 9 of 24 patients during the second follow-up examination, but only 2 of 15 during the third follow-up examination, and 0 of 4 during the fourth follow-up examination. Five repeat examinations were scheduled because of known residual polyps awaiting removal and one repeat examination was scheduled because of persistent protein-losing enteropathy. Polyp recurrence was investigated in 47 patients in whom all polyps were completely removed or eradicated during the initial

3 September 2010 JUVENILE POLYPS 797 direct sequence analysis and several also included testing for large-segment DNA deletion or duplication. Most patients were analyzed for mutations in SMAD4 and BMPR1A. Fewer patients were analyzed for mutations in PTEN, especially when phenotypic abnormalities suggestive for Bannayan Riley Ruvalcaba syndrome (also called PTEN hamartoma tumor syndrome) were noted. Abnormalities were detected in 10 patients involving SMAD4 (n 5), BMPR1A (n 2), and PTEN (n 3). This included 6 of the 9 patients with dysplasia: 2 with mutations in SMAD4 and 1 with a mutation in BMPR1A. In one case each involving BMPR1A and PTEN, the detected abnormality was a novel variant with as yet unproven pathologic significance. The patient with adenocarcinoma was not tested for genetic mutations but had features of hereditary hemorrhagic telangiectasia along with multiple juvenile polyps, findings that also were reported in the patient s father, who had a deletion in the SMAD4 gene, and paternal grandmother and was the subject of a prior case report. 22 Figure 1. All patients returning for repeat surveillance colonoscopy after index complete colonoscopy. Included were all colonoscopy examinations, both partial and complete. Treatment outcome and median time intervals between examinations are shown. complete colonoscopy (Figure 2). Initial repeat surveillance colonoscopy detected recurrence in 21 (44.7%) patients including 3 (16.7%) of 18 patients with a single polyp and 18 (62.1%) of 29 patients with multiple polyps noted at the time of index colonoscopy after a median interval of 15.1 months (interquartile range, mo). The number of patients in this subcohort who returned for repeat colonoscopy diminished over time with no patients returning for more than 4 follow-up examinations. Among the 3 patients with recurrence after eradication of a single polyp, 5 polyps were found in 1 patient 5 years later. In the second patient, 1 polyp was found 4 months later followed by 2 additional polyps 8 months after the first surveillance. A single polyp was found in the third patient after an interval of 16 months. Neoplasia Neoplasia was found in 10 (3.9%) of 257 patients with a mean age of 9.6 years (range, y). Seven of these 10 patients were boys. Low-grade dysplasia was found in 9 patients and adenocarcinoma with local lymph node metastases was found in one 11.8-year-old boy who had undergone multiple colonoscopy procedures without prior detection of dysplasia. Each of these patients with neoplasia had more than 5 polyps detected during the initial colonoscopy with a range of 7 to more than 100. The number of polyps that contained dysplasia included 1 polyp in 5 patients, 2 polyps in 3 patients, and more than 2 polyps in a single patient who had more than 100 juvenile polyps. Neoplasia involved the right colon in 7 of 10 patients, including the single case of adenocarcinoma. In 5 of these 7 patients, the neoplasia was found exclusively in the right colon. Genetics Genomic DNA from 23 patients with multiple polyps was tested for mutations associated with JPS. All included Discussion This was a large and comprehensive study on juvenile polyps in young patients. Although JPS is considered a rare diagnosis in the adult population, the finding of multiple juvenile polyps with numbers meeting the generally accepted clinical criteria of JPS is quite common in children and represented nearly one third of our cohort. Whether this represents the same population of patients found in adult series is unknown because the natural history of this disorder arising in childhood has not been studied carefully. Spontaneous sloughing of juvenile polyps, conversion to predominantly adenomatous lesions, and eradication of childhood juvenile polyps by the modern practice of colonoscopy with polypectomy are all possible explanations for the relative rarity of this polyp type in adult patients. Figure 2. Polyp recurrence in patients returning for repeat surveillance colonoscopy after complete polyp removal or eradication during the preceding examination. Only patients with complete colonoscopy examinations are included. Median time intervals between successive examinations are shown.

4 798 FOX ET AL CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 8, No. 9 Only 59% of patients in this cohort with multiple polyps underwent repeat colonoscopy despite published recommendations for continued surveillance colonoscopy in patients believed to be at risk for future neoplasia. 23 The variability we observed in the practice of repeat surveillance colonoscopy likely is representative of clinical practice in other pediatric referral centers during the same time period. Patients with fewer than 3 polyps generally have been considered low risk for polyp recurrence. No evidence-based pediatric practice guidelines have yet emerged. 24 Our findings, however, suggest that children who present with even a single polyp may continue to form polyps over time. This observation challenges the reliability of distinguishing between the diagnoses of solitary juvenile polyp and JPS on the basis of polyp number alone, a convention frequently applied when associated germline DNA mutations are not found or are not available and the family history is negative. Based on these findings, we suggest that a consistent practice of surveillance colonoscopy with polyp eradication in children presenting with one or more juvenile polyps is essential to determine a patient-specific profile of polyp recurrence and minimize the risk of developing advanced neoplasia. Although polyp recurrence is not uncommon within 1 to 2 years, the risk of developing advanced neoplasia in childhood remains low. Therefore, we recommend a starting interval of 1 to 3 years after initial eradication adjusting for low or high polyp burden and the presence of dysplasia during the index colonoscopy. Recommendations regarding the interval and duration of continuing surveillance examinations should rely on additional outcome data and on specific risk factors found in individual patients. We also emphasize the importance of complete colonoscopy in children as illustrated by our finding of exclusive right-colon involvement in 4.2% of cases and, more importantly, the predominant right colon location of neoplastic changes. Prophylactic colectomy may be offered to patients with a large polyp burden that precludes effective endoscopic management, one or more polyps with high-grade dysplasia, or a strong family history of early colorectal cancer associated with juvenile polyps. A nearly 4% rate of neoplasia in our study is comparable with rates reported in several pediatric series including 3 series with more than 100 patients. 2,3,20,25,26 However, the predominant right-colon location has not been noted in other reports. Dysplasia was found in 2 children ages 7 months and 12 years among the 137 patients reported by Dajani and Kamal 20 in Both cases involved solitary polyps and one included a focus of high-grade dysplasia. Poddar et al 2 in 1998 reported 17 (12%) cases of dysplasia (all low-grade) in 142 patients with juvenile polyps. Five of these 17 cases involved a solitary polyp. In 2001, Gupta et al 25 reported dysplasia in only 1 of 184 children with juvenile polyps. The patient was a 12-year-old boy with a single polyp containing a focus of high-grade dysplasia described as adenocarcinoma in situ. These findings taken together with other case reports of advanced neoplasia arising within a solitary juvenile polyp 10 suggest that children who form even small numbers of juvenile polyps are at increased risk for neoplasia including adenocarcinoma. Furthermore, the presentation of advanced adenocarcinoma in one young patient in our series who had no heralding dysplasia despite multiple prior examinations is sobering and illustrates the potential for insidious malignant transformation in this disorder. This particular patient had a combined phenotype of JPS and hereditary hemorrhagic telangiectasia, which has been reported in patients carrying germline mutations in SMAD4. 27 Retrospective reviews of adult patients with multiple juvenile polyps indicate much higher rates of neoplasia and adenocarcinoma than found in pediatric series. Giardiello et al 13 found neoplasia in 10 (18%) of 57 patients (mean age, 22 y) with one or more juvenile polyp. The mean age of patients with neoplasia was 37 years. This group included 3 adult patients with adenocarcinoma. Neoplasia was found in 9 of 19 patients with 3 or more polyps and in 1 of 28 patients with 1 or 2 polyps. Although only 23 patients in our series with multiple polyps underwent DNA analysis, a positive yield of approximately 40% is comparable with results recently reported by van Hattem et al 28 who performed a comprehensive analysis of SMAD4, BMPR1A, and PTEN in a series of patients with JPS using direct sequencing and multiplex ligation-dependent probe amplification. Our study also reports results of genetic mutation analysis in a cohort of exclusively pediatric patients. Our study had several limitations including its retrospective design and lack of uniform patient history, endoscopic and histologic reporting, and patient management protocols. Multiple clinicians were involved in the care of these patients and exercised independent clinical decision making in the absence of standardized treatment guidelines. Large-scale, multicenter, prospective studies would be useful to provide more accurate recurrence and long-term outcome data for children with juvenile polyps. Identification of biomarkers sampled from blood or urine 29 might guide the timing for and reduce the frequency of repeat colonoscopy, thereby reducing cost and enhancing quality of care. In conclusion, this large retrospective study of children with juvenile polyps provides new observations including a high rate of recurrence in patients with multiple polyps, a measurable rate of recurrence in patients presenting with a solitary polyp, a predominant right-colon location for neoplasia, and frequent association with identifiable gene mutations. Based on these data, we recommend the development of standardized protocols for repeat surveillance colonoscopy with polypectomy and genetic testing in children presenting with juvenile polyps. Supplementary Material Note: To access the supplementary material accompanying this article, visit the online version of Clinical Gastroenterology and Hepatology at and at doi: / j.cgh References 1. Horrilleno EG, Eckert C, Ackerman LV. Polyps of the rectum and colon in children. Cancer 1957;10: Poddar U, Thapa BR, Vaiphei K, et al. Colonic polyps: experience of 236 Indian children. Am J Gastroenterol 1998;93: Hoffenberg EJ, Sauaia A, Maltzman T, et al. Symptomatic colonic polyps in childhood: not so benign. J Pediatr Gastroenterol Nutr 1999;28: Turell R, De Maynard AL. Adenomas of the rectum and colon in juvenile patients. J Am Med Assoc 1956;161: Schapiro S. The occurrence of proctologic disorders in infancy and childhood; a statistical review of 2700 cases. Gastroenterology 1950;15: Gelb AM, Minkowitz S, Tresser M. Rectal and colonic polyps occurring in young people. N Y State J Med 1962;62:

5 September 2010 JUVENILE POLYPS Helwig EB. Adenomas of the large intestine in children. Am J Dis Child 1946;72: Smilow PC, Pryor CA Jr, Swinton NW. Juvenile polyposis coli: a report of three patients in three generations of one family. Dis Colon Rectum 1966;9: Stemper TJ, Kent TH, Summers RW. Juvenile polyposis and gastrointestinal carcinoma. A study of a kindred. Ann Intern Med 1975;83: Liu TH, Chen MC, Tseng HC, et al. Malignant change of juvenile polyp of colon: a case report. Chin Med J (Engl) 1978;4: Rolles CJ. Juvenile intestinal polyps are they always benign? Br Med J 1987;294: Jass JR, Williams CB, Bussey HJ, et al. Juvenile polyposis a precancerous condition. Histopathology 1988;13: Giardiello FM, Hamilton SR, Kern SE, et al. Colorectal neoplasia in juvenile polyposis or juvenile polyps. Arch Dis Child 1991;66: McColl I, Bussey HJR, Veale AMC, et al. Juvenile polyposis coli. Proc R Soc Med 1964;57: Nugent KP, Talbot IC, Hodgson SV, et al. Solitary juvenile polyps: not a marker for subsequent malignancy. Gastroenterology 1993;105: Howe JR, Roth S, Ringold JC, et al. Mutations in the SMAD4/ DPC4 gene in juvenile polyposis. Science 1998;280: Howe JR, Bair JL, Sayed MG, et al. Germline mutations of the gene encoding bone morphogenetic protein receptor 1A in juvenile polyposis. Nat Genet 2001;28: Sweet K, Willis J, Zhou XP, et al. Molecular classification of patients with unexplained hamartomatous and hyperplastic polyposis. JAMA 2005;294: Brosens LA, van Hattem A, Hylind LM, et al. Risk of colorectal cancer in juvenile polyposis. Gut 2007;56: Dajani YF, Kamal MF. Colorectal juvenile polyps: an epidemiological and histopathological study of 144 cases in Jordanians. Histopathology 1984;8: Gupta S. Clinical quiz. J Pediatr Gastroenterol Nutr 1998;26:174, Bross DA, Perez-Atayde A, Mandell VS, et al. Hereditary hemorrhagic telangiectasia presenting in early childhood. J Pediatr Gastroenterol Nutr 1994;18: Howe JR, Ringold JC, Hughes JH, et al. Direct genetic testing for Smad4 mutations in patients at risk for juvenile polyposis. Surgery 1999;126: Barnard J. Screening and surveillance recommendations for pediatric gastrointestinal polyposis syndromes. J Pediatr Gastroenterol Nutr 2009;48(Suppl 2):S75 S Gupta SK, Fitzgerald JF, Croffie JM, et al. Experience with juvenile polyps in North American children: the need for pancolonoscopy. Am J Gastroenterol 2001;96: Elitsur Y, Teitelbaum JE, Rewalt M, et al. Clinical and endoscopic data in juvenile polyposis syndrome in preadolescent children: a multicenter experience from the United States. J Clin Gastroenterol 2009;43: Gallione CJ, Repetto GM, Legius E, et al. A combined syndrome of juvenile polyposis and hereditary haemorrhagic telangiectasia associated with mutations in MADH4 (SMAD4). Lancet 2004; 363: van Hattem WA, Brosens LA, de Leng WW, et al. Large genomic deletions of SMAD4, BMPR1A and PTEN in juvenile polyposis. Gut 2008;57: Manfredi MA, Fox VL, Zurakowski D, et al. Matrix metalloproteinases in the urine and tissue of patients with juvenile polyps: potential biomarkers for the presence of polyps. J Pediatr Gastroenterol Nutr 2009;48: Reprint requests Address requests for reprints to: Victor L. Fox, MD, Children s Hospital Boston, 300 Longwood Avenue, Boston, Massachusetts victor.fox@childrens.harvard.edu; fax: (617) Conflicts of interest The authors disclose no conflicts.