tension-type, migraine or both types of headache.

Results: To date, 288 subjects have completed follow-up including 88 former well controls and 200 former CAP patients. The mean age at follow-up was 20.2 years and 59% were female. Using chi-square analysis, we found that headaches were significantly more common in those with FGIDs (58%) compared to those without FGIDs (38%, p<0.05). Migraine headaches were also significantly more common with FGID's (45%) compared to those without FGIDs (30%, p<0.05). There was not an association between tension-type headaches and FGIDs (P>0.05). In evaluation of type of FGID and headache type, it appears that migraines are significantly more common with functional dyspepsia (51%) compared to those without functional dyspepsia (31%, p<0.05). Migraines were not statistically associated with other FGID types including abdominal migraine, IBS, FAP, and FAPS. However, there was a trend between migraines in those with IBS (45%) compared to those without IBS (32%, p=0.078). Conclusions: In our prospective cohort study, there is a statistically significant association between FGID and headache diagnosis. There also was significant association between migraines and both FGID and functional dyspepsia. There was a trend between migraines and IBS. Further research is needed to investigate the pathophysiology of these findings. 510 Long Term Prognosis of Children Undergoing HIDA Scan for Suspected Biliary Dyskinesia Sudipta Misra We have previously reported short term prognostic value of Hepatobiliary Scintiscan (HIDA) in children suspected to be having functional biliary tract disease. In this abstract we report 5 year follow up of the population. Methods: Children undergoing HIDA scan at Children's Hospital of Illinois for chronic abdominal pain, nausea or vomiting between March 1999 and February 2002 were analyzed. These children had inconclusive gastrointestinal diagnostic workup including hepatobiliary ultrasound, endoscopies, barium X rays, pH probe studies etc. Therapeutic trials in some patients with acid suppression, tricyclic antidepressants and anti spasmotics were not successful. HIDA scan was performed according to a standardized protocol. A gall bladder ejection fraction of less than 35% was taken to be abnormal. Clinical data was collected by retrospective chart review. A scripted telephone survey was done 5 years after the initial HIDA scan to document long-term outcome. The data was re analyzed to compare short and long term prognosis of these children. Results: Forty two of 61 children had abnormal HIDA scan. On telephone follow up at an average interval of 61.9 months after HIDA scan; 19 children (31.1%) had recurrence of symptoms, 33 (54.1%) were asymptomatic and 9 (14.8%) could not be reached There was no difference between children with normal and abnormal HIDA results in clinical presentations, short term (85.7% and 84.2%) and long term (64.9% and 60%) outcomes. Twenty seven of the 42 children with abnormal scan results underwent interventions (21 cholecystectomy only, 4 cholecystectomy followed by sphincter of Oddi sphincterotomy and 2 sphincterotomy only). Following intervention, children with abnormal HIDA scan had better short-term prognosis (88.9% and 54.5%, p=.02) but their long-term prognosis (52.2% and 85.7%, p = .04) was worse than those without intervention. No clinical prognostic indicator such as location of abdominal pain, nausea etc. could be identified. Conclusion: Children with abnormal HIDA scan had a high rate of long term recurrence after surgical intervention as well as spontaneous resolution of symptoms without intervention. This test was not a good long term prognostic indicator. Hence, HIDA scan result, including undetectable ejection fractions, should be used with caution to select children for surgical intervention in suspected biliary dyskinesia . Focused prospective studies are needed to define biliary dyskinesia in children. 511 Clinical Features Associated With GI Symptoms in Autism Spectrum Disorders (ASD) Kent C. Williams, George J. Fuchs, Glenn T. Furuta, Margaret A. Marcon, Daniel L. Coury Background: Rates of ASD are estimated at 1 in 91 children(1.1%). While studies report conflicting results about the prevalence of GI symptoms in children with ASD versus children without ASD, GI complaints are common in both populations. However, whether GI symptoms in children with ASD are associated with distinctive clinical features is unclear. Objectives: To determine the frequency of GI symptoms as reported by parents in a large ASD registry, and to identify clinical features associated with GI symptoms in children with ASD. Methods: Autism Treatment Network Registry enrolled 1420 children with an ADOSconfirmed ASD diagnosis (autism, Asperger disorder, or PDD-NOS) between September 2007 and December 2009 at 15 sites in US and Canada. Parents completed a GI symptom inventory, as well as Child Behavior Checklist (CBCL), Child Sleep Health Questionnaire (CSHQ) and Pediatric Quality of Life (PedsQL) at time of enrollment. Results: GI data was available for 1185 children. At time of enrollment 45% of children displayed GI symptoms. For GI complaints present within the previous 3 months, constipation was most common (32%) followed by abdominal pain (27%), diarrhea (26%), other (18%), nausea (14%) and bloating (12%). For chronic GI complaints (>3 months duration), constipation was most common (22%) followed by other (14%), abdominal pain (14%), diarrhea (12%), bloating (9%) and nausea (5%). GI symptoms increased with age, ranging from 39% in those under 5 years to 51% in those 7 years and older (p<0.0001). Children ages 1 to 5 years with GI symptoms had higher CBCL t-scores for total problems and for the emotionally reactive, anxious/depressed, somatic complaints, sleep problems, internalizing problems, affective problems, and anxiety problems subscales, all p<0.05. Children age 6 to 18 years with GI symptoms had higher CBCL t-scores for total problems and for all subscales (p<0.01). Sleep problems occurred more frequently in children with than those without GI symptoms (50% versus 37%, p<0.0001). Children with GI symptoms also had lower PedsQL scores (overall score and all five subscales, p<0.01) compared to children without GI problems. Presence of GI problems did not differ by gender, ASD subtype, race, or IQ. Conclusions: Parents of children with ASD report a high prevalence of GI symptoms in their children. GI complaints are associated with age, behavioral abnormalities, sleep disturbances and overall decreased health-related quality of life. Further investigation is needed to clarify these associations and

whether treatment of GI disorders improves behavioral and cognitive function, sleep, and quality of life in ASD children. 543 Intestinal Wall Remodeling Measured With Serial CT Enterography in Patients With Crohn's Disease Treated With Infliximab David H. Bruining, Edward V. Loftus, Eric C. Ehman, Hassan A. Siddiki, Douglas L. Nguyen, Jeff L. Fidler, James E. Huprich, Jayawant N. Mandrekar, William S. Harmsen, William J. Sandborn, Joel G. Fletcher Background: The use of computed tomography enterography (CTE) in patients (pts) with Crohn's disease has increased at some tertiary care centers. CTE has a high sensitivity and specificity for active small bowel inflammation, and it may provide an objective measurement of treatment response. There is a paucity of data, however, how radiologic parameters of active disease change over time with infliximab therapy, and whether these alterations correspond to clinical symptoms, serum biomarkers, or endoscopic appearance. Methods: We retrospectively identified pts with established Crohn's disease who had undergone serial CTE imaging while receiving infliximab. Pts had imaging before and after initiation of infliximab, or two CTEs more than 6 months apart while on maintenance dosing. Data were analyzed both per lesion and per pt. Lesions were defined as improved if a decrease in enhancement or length occurred without worsening of other parameters (enhancement, length, vasa recta dilatation/comb sign, fatty proliferation, or stratification). Pts were grouped as responders (all lesions improved), partial responders (some lesions improved), and nonresponders (worsening or no changes in all lesions). In order to remain in the responder group, no worsening could be identified on CTEs subsequent to the initial follow-up examination. Radiologic scoring was performed by a GI radiologist blinded to the clinical information. Clinical symptoms at the time of follow-up CTE was determined by a gastroenterologist blinded to the imaging results. Results: Of the 63 pts identified (52% males), median age was 37.7 years, median disease duration was 14.9 years, and median time between initial and first follow-up CTE was 356 days. Only 17 patients (27.0%) did not have pre-infliximab CTE imaging for comparison. The most common indications for repeat imaging were reassessment of disease activity (79.3%) and pain (17.5%). A total of 105 lesions were identified, 52 (49.5%) improved, 11 (10.5%) remained unchanged, and 42 (40.0%) worsened. Per pt, 28 (44.4%) were responders, 12 (19.0%) were partial responders, and 23 (36.5%) were non-responders. Radiologic response showed a low agreement with clinical symptoms at time of 2nd CTE (Kappa=0.28), endoscopic appearance (Kappa=0.16), and CRP (Kappa= 0.39). Conclusion: Radiologic improvement/remodeling was noted in more than half of pts on infliximab therapy, despite a study design that was likely biased towards non-responders. A poor correlation with clinical symptoms, serum biomarkers, and endoscopic appearance supports the notion that CTE imaging may be a complementary tool to other methods in assessing response to treatment. 544 Predicting Response to Infliximab in Crohn's Disease: Role of CRP, MR Enterography and NOD2 Genotype Finlay A. Macrae, Gregor J. Brown, Paula Lewis, Bernadette Viney, Cathy Pizzey, Zina Valaydon, Damien Stella, Robert N. Gibson, Kaye E. Marion Response to infliximab therapy for Crohn's disease (CD) is unpredictable. We aimed to i) correlate MR enterography (MRE) findings with the CD Activity Index (CDAI) and C Reactive Protein (CRP) at study entry and exit, ii) determine if entry a) CRP, b) NOD2 genotyping and/or c) MRE could predict response to infliximab as measured by CDAI and MRE, and iii) correlate changes in MRE with changes in CRP and CDAI in response to treatment. Methods: 29 consenting patients with ileal CD, CDAI >220 and clinically assessed as requiring infliximab were recruited. NOD2 genotyping and, at entry and exit (at 26 weeks), MRE (the primary endpoint), CDAI, and CRP assessments were done. 24 completed the study; 4 withdrew for surgery and 1 moved overseas. Disease remission and response were defined as CDAI<151 and a reduction in CDAI>100 respectively at study end. 3 infliximab infusions were administered for induction. Relationships assessed were: CDAI with CRP at entry and exit; NOD2 mutational status with CDAI change during treatment; MRE parameters with CDAI and CRP at entry and exit; and entry MRE parameters (length of affected bowel; presence of strictures; wall thickness and contrast enhancement; and global MR assessment of disease activity) with changes in CDAI (remission and/or response) and CRP during treatment. Results: Of the 24 patients, 15 entered remission and 6 more responded. All but 6 patients normalized CRP with treatment. i) There was no correlation between the CDAI and CRP at entry or exit; iia) CRP at baseline did not predict remission or response to infliximab. iib) 5/21 had pathogenic mutations in NOD2; the median CDAI change for patients with mutations was 293 compared with 224 for those without mutations. iic) On global MRE assessment, 5 of 23 evaluable patients normalized findings at exit MRE, 5 improved and 13 had no change between entry and exit. Improvement in CDAI was greater for patients with fewer strictures at entry (p=0.04) whereas bowel wall thickening and length of bowel affected did not significantly predict response. iii) MRE global response correlated with changes in CDAI (p=0.04) but not CRP. Conclusion: As elsewhere, we found no correlation of CDAI with CRP. Patients with NOD2 mutations had a greater fall in CDAI than those without NOD2 mutations. Global MRE assessment of response correlated with CDAI changes; multiplicity of strictures predicted a poor response to infliximab. MRE parameters at exit identified continuing disease more sensitively than CDAI, supporting continuation of treatment beyond symptomatic response (i.e. CDAI).

AGA Abstracts

AGA Abstracts

S-74