Cohen, Stanley A.*; Ephrath, Hagit†; Lewis, Jeffery D.*; Klevens, Alan†; Bergwerk, Ari†; Liu, Steven*; Patel, Dinesh*; Reed-Knight, Bonney*; Stallworth, Angela*; Wakhisi, Tamara*; Gold, Benjamin D.*
Capsule endoscopy (CE) has been adopted widely as a clinical tool in the evaluation of small bowel (SB) disease for a variety of conditions, and it received marketing clearance by the US Food and Drug Administration in 2003 for CE in pediatric patients 10 years of age and older. Supported by additional experience in children as young as 10 months of age, in 2009 the Food and Drug Administration expanded the role for CE use to children 2 years and older and approved the use of a patency capsule (PC) for the same age group (1).
Because CE avoids ionizing radiation, deep sedation, or general anesthesia required by other diagnostic imaging or endoscopic methods, CE has the potential to be particularly valuable in pediatrics (2–19). Nevertheless, there are few large-scale reports on the broad use of CE in pediatric populations and a single small prospective trial of the PC. We therefore examined our center's cumulative experience to better understand the safety and efficacy of CE and the PC in children and how the PC affects the yield and safety of CE.
The present retrospective study was approved by the institutional review board at Children's Healthcare of Atlanta. All of the CE studies performed at the Children's Center for Digestive Health Care in Atlanta, GA, from July 2004 to June 2009 were serially registered. The relevant data in their capsule studies and patient medical records were abstracted and reviewed retrospectively. Demographic data, indications, and findings were recorded, along with frequency of capsule retention, rates of completion of the CE procedure, use of a PC, and findings of the patency and CE examinations.
SB CE examinations (PillCam SB, Given Imaging, Yokneam, Israel) were performed according to the manufacturer's recommendation, with a10-hour fast before ingestion of the capsule. Simethicone 60 to 160 mg was given at the time of ingestion, with no other prokinetics or laxative cleanout used. The patient either swallowed the PillCam SB capsule or the capsule was placed via upper endoscopy into the duodenum or the stomach, while the child was under general anesthesia. The latter was accomplished with a CE delivery device (AdvanCE, US Endoscopy, Mentor, OH) or Roth Net retrieval device (US Endoscopy). The resulting videos were interpreted by 1 of 4 physicians who were trained at the American Society for Gastrointestinal Endoscopy, attended Given Imaging–sponsored courses, or trained by a primary mentor during fellowship. When the CE reading was uncertain, the initial reader had the option of obtaining a corroborative evaluation from one of the other physicians.
A commercially available, lactose- and barium-filled dissolving PC developed to test functional intestinal patency (Agile Patency System, Given Imaging) was administered at the discretion of the treating physician. Those patients who then passed an intact PC, usually within 40 hours, were able to ingest the standard SB capsule (according to the standard protocol above) within a 2-week time frame.
Summary statistics (arithmetic mean, standard deviation, minimum and maximum for quantitative variables) were determined for subject characteristics. Capsule diagnostic yield was calculated based on observed pathologies. Pathologies were categorized according to their relation to the patients’ underlying disease and location. Disease categories were defined as follows: confirmed CD, suspected CD, occult gastrointestinal (GI) bleeding or anemia, polyposis, confirmed celiac, and inconclusive or normal SB. GI locations were defined as gastric, SB, and colon. Diagnostic yield was reported as raw numbers and as proportions.
Capsule findings were compared against subjects’ indication for CE referral by the treating physician. Change in terms of establishing present diagnosis and detection of active disease was reported as raw numbers and as proportions. No statistical comparisons were performed because of the variability in the patient evaluation before the CE procedure.
Completion of CE procedure was defined as cecum visualization and was reported as raw numbers and as proportions. In addition, effectiveness of the incomplete procedure with respect to clinical effect was analyzed in a qualitative manner.
A total of 284 CE and 23 patency procedures were performed in 277 patients during the 5 years reviewed. One patient underwent 4 CE procedures within 4 years; 3 patients have undergone 2 procedures within 2½ years, with an additional 2 patients having repeat procedures within 6 weeks because of the limited nature of the initial procedure. Mean age was 15 years (±3.7), with a range of 3.4 to 23 years.
Twenty-eight children younger than 10 years of age (22 boys) had 31 CE, with 1 having a PC before CE. Eleven capsules were swallowed, whereas 20 were delivered endoscopically into the duodenum or stomach. The average age of this group was 7.2 ± 1.8 years; the youngest was 3.4 years (weight 13.5 kg). The youngest to independently swallow the capsule was 4.6 years. No capsule retentions occurred in this group; however, 9 CE had incomplete studies: 2 had malfunctions, 7 did not exit the SB during CE (3.7- to 8.0-hour SB time), and in 3, debris obscured substantial amounts of the SB. There were no significant differences in age, height, weight, body mass index, hemoglobin, or albumin in those who had incomplete or complete studies. Gastric passage was 75.3 ± 65.7 minutes for those who swallowed the capsule. SB passage was similar for those who swallowed the capsule versus those for whom it was placed (3.26 ± 1.0 vs 4.0 ± 1.3 hour).
The indications for CE are listed in Table 1. Two hundred forty-five (86%) CE studies were performed for suspected (184, 65%) or confirmed (61, 21%) CD. Anemia or obscure GI bleeding was the indication in 27 (9.5%) procedures and 6 (2%) were done for polyposis. CE was performed in 4 procedures (1.4%) to evaluate celiac disease and in 2 others, 1 with lymphangiectasia and 1 with dermatomyositis presenting with nausea, abdominal pain, and weight loss.
Positive findings were found in 205 (72%) of the studies, with 152 (54%) having SB findings. Of these, 72 (47%) were diagnostic, whereas 80 (53%) had findings that were less specific. Gastric abnormalities were seen on 95 (33%) procedures and colonic abnormalities were identified on 31 (11%) of the total procedures, including 50 gastric and 17 colonic abnormalities when the SB was considered normal (Table 2).
In patients younger than 10 years of age, 23 CE were performed for suspected CD (6 CD cases were found, 3 having had previous negative colonoscopy, 1 previously identified proctitis, 1 previously identified mild colitis); 8 had CE for other reasons (6 for bleeding, 1 known CD, 1 lymphangiectasia). These demonstrated findings in 5: vascular lesions in 3, lymphangiectasia 1, and CD 1. Positive findings were present in 10 patients with completed studies and 1 patient with an incomplete study.
Of the 245 CE studies in known or suspected CD, 63 (26%) identified small intestinal lesions that were consistent with inflammatory bowel disease (IBD), whereas 38 (16%) studies were suggestive of CD but not diagnostic, according to commonly accepted criteria (20), and another 33 demonstrated nonspecific findings in the SB, that is, nondiagnostic lesions such as erythema, edema, or mucosal irregularities. Eighty-nine studies showed gastric ulcers or varying degrees of gastritis and severe colitis on 2 CE procedures in which the SB was normal.
Of 184 patients who were referred for CE as suspected CD because of active symptoms, CE identified inflammatory lesions and/or at least 3 ulcers that led to the diagnosis of CD in 28 (15%) patients (Figs. 1 and 2). In 56 (30%) patients, CE identified inflammatory lesions, but the findings were not sufficient to establish CD based on CE findings alone. In 2 (2%) of these patients, CE identified active bleeding. A normal SB was found in 99 (54%) patients. Seventy-four (40%) patients had gastric abnormalities, including 1 gastric polyp among the 28 patients who had gastric lesions, whereas the SB was normal. In the 14 patients classified as suspected CD in which pain was the only symptom other than nausea or previous gastric findings (ulcer in 1, gastritis in 1) (Table 2), CD was found in only 2 patients, but other findings included severe, distal lymphangiectasia (1), nonspecific edema and erythema (1), gastric ulcers (3), gastritis (1), and colitis (2), whereas 3 studies were entirely normal.
Of the 48 patients who had suspected CD with previously diagnosed colitis, CE identified inflammatory lesions in the SB and led to the diagnosis of SB CD in 7 (15%) patients. In 1 case, colonic involvement was visualized within the CE procedure. In 8 (17%) patients, CE identified inflammatory lesions in the SB, but the findings were not sufficient enough to establish SB CD. Normal SB or insignificant lesions that could not lead to conclusive diagnosis in the SB were present in 33 (69%) patients.
Among the 13 patients with active GI bleeding and 14 with fecal occult blood or anemia, SB findings were present in 6 (46%) and 4 (29%) for a total of 10 of 27 (37%), with CD found or strongly suspected in 3 and 1 in the groups, respectively. Polypoid lesions were seen in 3 patients, a vascular lesion was found in 1, and duodenal and gastric lesions were present in 2 and 4 patients, respectively. Colonic abnormalities were found in 3 patients, including colonic ulcerations in 1 patient who also had gastritis.
In the 6 patients referred with various polypoid conditions (specific diagnoses of the patients referred included Peutz Jeghers, neurofibromatosis, familial polyposis, multiple gastric polyps, and multiple colonic polyps), 4 demonstrated polyps, although 2 were isolated to either the stomach or colon; 1 had significant lymphoid hyperplasia; and 1 was entirely normal. In the 4 patients evaluated for celiac disease, 2 had erythema and apparent loss of villi with 1 showing significant scalloping (Fig. 3), whereas 2 were visually normal. The patient with known lymphangiectasia (as well as the patient described above with lymphangiectasia at CE) had predominantly distal disease (Fig. 4) with sparing of the duodenum. The patient with dermatomyositis had nonspecific erythema in the duodenum and ileocecal area.
Completion and Retention
In 65 (23%) CE procedures, the capsule did not reach the cecum during the study and were considered incomplete. These include 4 patients in whom the capsule remained in the stomach throughout the recording time but eventually passed. In the remaining 61 studies, small intestinal lesions were found in 35 (58%) patients, including subsequent confirmation of CD in 26 (43%) patients. Patients and parents reported eventual passage of the capsules. Where they were uncertain whether the capsule had passed, an abdominal x-ray was obtained.
Retention of the capsule, defined as nonpassage of the capsule within 14 days or removal of the capsule before that time, was observed in 7 (2.4%) studies (6 with known CD, 1 with suspected CD). Six required surgical intervention to retrieve the capsule, including 1 patient in whom the capsule was administered to identify the location of the stricture to guide the surgeon, and another intervention in which CE was performed postoperatively in the same patient and the capsule was again retained. One patient underwent endoscopic removal of a retained capsule through a tight ileostomy.
A total of 23 patients underwent evaluation with the PC before using the video capsule in patients with known (14) or suspected (9) CD. Two of the latter group had colitis (1 ulcerative colitis [UC], 1 indeterminate colitis [IC]). In 3 of these patients (all with known CD), this procedure did not demonstrate intestinal patency and the video capsule was not offered. One patient successfully passed the PC at 24 hours but did not return for his CE and was lost to follow-up. In the 19 others, patency was established and a subsequent video capsule study was performed. In these, CD was confirmed in 10 and suspected in 1, and gastric lesions were identified in 4. Normal or inconclusive small intestinal findings were seen in 3 patients.
In 1 patient in whom patency was established 1 week before CE, the video capsule remained in the SB at the end of the study (patient is included in those with a retained capsule above). The results of the study led to the diagnosis of CD and the patient subsequently underwent a needed ileocecal resection.
The present retrospective review of the CE, performed during a 5-year period, characterizes the largest pediatric cohort published in the literature to date. In addition, this report describes the first clinical use of a PC in pediatrics, other than in a protocolized, prospective study. Our review found CE to be especially useful in the evaluation of IBD and the PC helpful in lessening the potential for capsule retention. Moreover, the number of gastric and colonic findings argue for a complete reading of capsule studies rather than SB-focused evaluations that are abbreviated and more time efficient. Based on these findings and the less invasive nature of CE, an additional possibility is that CE should become the first endoscopic procedure considered, instead of traditional endoscopy, when evaluating the intraluminal cause of anemia or bleeding or when monitoring the patient with known or suspected CD.
Similar to other studies in the pediatric age group (2–15,18), a much higher proportion of CE procedures (86%) were performed in our retrospective cohort to evaluate patients with known or suspected CD than in adult series (19), in which anemia and or occult GI bleeding are the predominant indications for CE (Table 1). Twenty-eight (10%) of our patients were the focus of an earlier retrospective review of CE in patients with known IBD (7). Of that group, 4 of 5 patients with UC and 1 of 2 patients with IC (total 5/7, 71% of UC/IC patients) had their disease reclassified to CD based upon newly diagnosed SB mucosal lesions. In addition, 13 of 21 (62%) patients with CD had more extensive SB disease, including newly detected jejeunal disease (12/13; 92%) at the time of CE examination. In the 5 newly diagnosed patients with CD, all had therapeutic changes made following the findings at CE.
This early experience with CE in helping to tailor optimal management of the disease may have led to the high degree of CE use for the evaluation of patients with IBD. Additionally, patients with isolated, persistent, or unexplained pain in our cohort have fallen under the classification of suspected CD to meet standardized guidelines for CE and obtain appropriate insurance reimbursement (because they could be suspected to have CD on the basis of pain alone). Previous studies have separated these patients into another category (Table 1).
Our detection rate was higher than in other studies because we included our findings in the stomach and colon; however, our findings in the SB (54%) were similar to other studies and showed little variability for our patients with IBD (55%) and our overall diagnostic yield in non-IBD patients at 56% (22/39) with 19 of 29 (48%) present in the small intestine. Some of our patients, particularly those with known IBD, may have benefited even when no SB involvement was found (although this is deemed a negative result by standard methodology (20)) because a CE procedure may have demonstrated the lack of small intestinal involvement in IC or mucosal improvement despite active symptoms, thereby helping to direct further management.
The rate of incomplete CE studies reported here (23%) may be a reflection of the use of general anesthesia at the time of endoscopic placement in many of our younger patients or because many of the capsules used in our series were first-generation video capsules, which had a shorter battery life than the presently available SB2 capsule. Even when the studies were incomplete, small intestinal disease could be identified in 58%, resulting in a diagnosis of CD in 43%.
The common and valid concern regarding capsule retention may have been partially ameliorated in our cohort by using the PC in high-risk patients (usually those with known or suspected CD with possibly mild obstructive symptoms, such as nausea). We found only 5 cases of retention (1.8%) in nonsurgical patients despite the high prevalence of IBD in our series. Three CE procedures were cancelled when patients failed the PC procedure. Our experience in the present retrospective study parallels our previously reported experience in a separate prospective cohort testing the timing and use of the PC in small intestinal IBD (21). In this previously reported prospective study, 18 patients (10–16 years of age; 9 boys; 5 known CD, 3 IC, 1 UC, 9 suspected CD) ingested the PC. Fifteen patients excreted intact PC (mean 34.5 hours), 12 patients within 40 hours (range 9–60 hours). Sixteen (89%) patients underwent subsequent CE successfully. There were no capsule retentions or adverse events (21); however, 1 patient in the present retrospective cohort did undergo the PC procedure but had retention of a video capsule 1 week after the PC passed uneventfully within 30 hours. This may suggest that the PC procedure demands careful interpretation, and patients with CD should be aware of the higher risk of retention under any circumstances. Of note, no factors other than the presence of known CD (6/7) were predictive of retention, and no established protocol led to the ordering of a PC before CE, although the findings here suggest that PC should be obtained in patients with known CD undergoing CE.
The principal limitation of the present study in describing the full range of CE utility is emphasized by the predominance of patients with known and suspected CD. There were a much smaller number of patients in which CE was performed for other indications, thereby making comparisons between indications difficult. This large population of IBD does, however, allow us to recognize the potential for evaluating mucosal change with therapy and the range of findings that may be found in those presenting with abdominal pain. At the same time, the experience presented here understates other aspects of CE that may be better understood from previously published systematic reviews and multicenter studies that need to be done. Additionally, the lack of consistency with regard to the patient workup before the CE procedure, the lack of validated criteria for diagnosing mucosal disease, the lack of tissue sampling of suspected CE pathologies to provide a criterion standard for comparison, and the potentially different diagnostic thresholds by the different readers affect diagnostic sensitivity and specificity. Our high proportion of gastric findings, many of them nonspecific, does raise some question of their significance and clinical relevance, especially because our retrospective review did not seek to provide correlation with biopsy results. Future studies focusing on these issues and ones that further improve diagnostic accuracy are needed to realize the full potential of CE in pediatrics.
This large, 5-year retrospective series confirms the validity of the CE as a diagnostic tool in the evaluation of SB disease in pediatrics, especially in those patients with IBD. Our study also demonstrated that this diagnostic device can be safe and effective even in small-sized pediatric patients, that is, a patient with a weight of 13.5 kg. The CE reader must be aware that pertinent findings, with diagnostic and therapeutic implications, may be apparent beyond the SB.
Furthermore, our study demonstrated that there remains a relatively small risk of capsule retention, particularly in the patient with known or suspected IBD. Moreover, we showed that the PC may be helpful in minimizing the risk of capsule retention. The rate of incomplete studies using CE may be more frequent in pediatric patients, but even under such circumstances, useful diagnostic information can be obtained.
1. US Food and Drug Administration Center for Devices and Radiological Health. PC Patency System and Pillcam Platform With Pillcam SB Capsules. (510k Number K090557; Approval September 28, 2009) Available at: http://www.accessdata.fda.gov/cdrh_docs/pdf9/K090557.pdf
. Accessed February 10, 2010.
2. Jensen MK, Tipnis NA Bajonrunaite R, et al. Capsule endoscopy performed across the pediatric age range: indications, incomplete studies and utility in management of inflammatory bowel disease. Gastrointest Endosc 2010;72:95–102.
3. Thomsom M, Fritscher-Raven A, Mylonaki M, et al. Wireless capsule endoscopy in children: a study to assess diagnostic yield in small bowel disease in pediatric patients. J Pediatr Gastroenterol Nutr 2007; 44:192–197.
4. Tokuhara D, Watanabe K, Okano Y, et al. Wireless capsule endoscopy in pediatric patients: the first series from Japan. J Gastroenterol 2010;45:683–91.
5. Guilhon de Araujo Sant’Anna AM, Dubois J, Miron MC, et al. Wireless capsule endoscopy for obscure small bowel disorders: final results of the first pediatric controlled trial. Clin Gastroenterol Hepatol 2005; 3:264–270.
6. Antao B, Bishop J, Shawis R, et al. Clilnical application and diagnostic yield of wireless capsule endoscopy in children. J Laparoendosc Adv Surg Tech 2007; 17:364–370.
7. Cohen SA, Gralnek IM, Ephrath H, et al. Capsule endoscopy may reclassify pediatric inflammatory bowel disease: a historical analysis. J Pediatr Gastroenterol Nutr 2008; 47:31–36.
8. de’ Angelis GL, Fornaroli F, de’ Angeles N, et al. Wireless capsule endoscopy for pediatric small-bowel diseases. Am J Gastroenterol 2007; 102:1749–1757.
9. Atay O, Mahajan L, Kay M, et al. Risk of capsule endoscope retention in pediatric patients: a large single-center experience and review of the literature. J Pediatr Gastroenterol Nutr 2009; 49:1–6.
10. Moy L, Levine J. Capsule endoscopy in the evaluation of patients with unexplained growth failure. J Pediatr Gastroenterol Nutr 2009;48:647–50.
11. Moy L, Levine J. Wireless capsule endoscopy in the pediatric age group: experience and complications. J Pediatr Gastroenterol Nutr 2007; 44:516–520.
12. Ge ZZ, Chen HY, Gao YJ, et al. Clinical application of wireless capsule endoscopy in pediatric patients for suspected small bowel diseases. Eur J Pediatr 2007; 166:825–829.
13. Arguelles-Arias F, Caunedo A, Romero J, et al. The value of capsule endoscopy in pediatric patients with a suspicion of Crohn's disease. Endoscopy 2004; 36:869–873.
14. Urbain D, Tresinie M, De Looze D, et al. Capsule endoscopy in paediatrics: multicentric Belgian study. Acta Gastroenterol Belg 2007;70:11–4.
15. Fritscher-Ravens A, Scherbakov P, Bufler P, et al. The feasibility of wireless capsule endoscopy in detecting small intestinal pathology in children under the age of 8 years: a multicentre European study. Gut 2009;58:1467–72
16. Shamir R, Hino B, Hartman C, et al. Wireless video capsule in pediatric patients with functional abdominal pain. J Pediatr Gastroenterol Nutr 2007; 44:45–50.
17. Postgate A, Hyer W, Phillips R, et al. Feasibility of video capsule endoscopy in the management of children with Peutz-Jeghers syndrome: a blinded comparison with barium enterography for the detection of small bowel polyps. J Pediatr Gastroenterol Nutr 2009; 49:417–423.
18. Cohen SA, Klevens AI. Use of capsule endoscopy in diagnosis and management of pediatric patients, based on meta-analysis. Clin Gastroenterol Hepatol 2011;9:490–6.
19. Liao Z, Gao R, Xu C, et al. Indications and detection, completion and retention rates of small bowel capsule endoscopy: a systematic review. Gastrointest Endosc 2010; 71:280–286.
20. Triester SL, Leighton JA, Leontiadis GI, et al. A meta-analysis of the yield of capsule endoscopy compared to other diagnostic modalities in patients with obscure gastrointestinal bleeding. Am J Gastroenterol 2005; 100:2407–2418.
21. Cohen SA, Gralnek I, Ephrath H, et al. The use of patency capsule endoscopy in pediatric Crohn's disease—a prospective evaluation. Digest Dis Sci 2011; 56:860–865.