Gastroesophageal reflux in children older than years

Transcription

1 CLINICAL GASTROENTEROLOGY AND HEPATOLOGY 2007;5: Determinants of Gastroesophageal Reflux Disease in Adults With a History of Childhood Gastroesophageal Reflux Disease HASHEM B. EL SERAG,*, PETER RICHARDSON,* PETRA PILGRIM,* and MARK A. GILGER *Sections of Health Services Research and Gastroenterology, Michael E. DeBakey Veterans Affairs Medical Center, and Department of Pediatrics, Baylor College of Medicine, Houston, Texas Background & Aims: We conducted a nested case-control study to examine the prevalence and risk factors for current gastroesophageal reflux disease (GERD) symptoms in young adults with a history of childhood GERD. Methods: We identified a cohort of individuals diagnosed with GERD in childhood during , and controls without childhood GERD. Patients with neurodevelopmental disorders, tracheoesophageal anomalies, or cystic fibrosis were excluded. A computer-assisted telephone interview was conducted during We calculated the prevalence of GERD symptoms, and examined the potential determinants of symptoms in unadjusted and adjusted logistic regression analyses. Results: A total of 113 cases completed the questionnaires (participation rate, 70.6%). The mean age of participants was 18 years, and their mean age at the time of childhood GERD diagnosis was 10 years. At least weekly heartburn or regurgitation was reported in 52 (46%) participants, 94% of whom were taking proton pump inhibitors, H2RA, or antacids. On the other hand, 33 controls were identified (44% participation rate) in whom weekly heartburn or regurgitation was reported in 30%. GERD was significantly more frequent in females using oral contraceptive pills (76.5%) as compared with females not on oral contraceptive pills (47.9%), or males (33.3%) (P.008). Weight, height, nonsteroidal anti-inflammatory drug use, race, family history of GERD, education level, employment status, tobacco smoking, alcohol, or coffee drinking were not associated significantly with adulthood GERD. Conclusions: Frequent GERD symptoms requiring antisecretory therapy are present in approximately half of young adults with a history of childhood GERD. The use of oral contraceptives is a risk factor for GERD symptoms in these individuals. Gastroesophageal reflux in children older than years of age can persist and become severe, producing esophageal or extraesophageal damage, and is therefore considered gastroesophageal reflux disease (GERD). 1,2 Childhood GERD is associated with neurologic disorders (eg, cerebral palsy), congenital malformations (eg, esophageal atresia 3 ), and cystic fibrosis. However, a large proportion (40% 60%) of children with GERD have none of these underlying disorders. The clinical course of GERD in that latter group of children is unclear. Most previous studies that have evaluated the clinical course of children with GERD examined cases with infantile GERD, and/or examined a mix of patients with or without severe underlying illness. 4 6 However, 1 retrospective study of 76 children with GERD and no comorbid illness reported spontaneous resolution of symptoms in only a quarter of these patients after an average of 28 months of follow-up evaluation. 7 Another small retrospective study of 32 neurologically normal children with GERD reported that after a period of 3.4 years, spontaneous resolution of symptoms occurred in 41%, symptom improvement with medical therapy in another 41%, and persistent symptoms in 13% despite medical therapy. 2 These 2 studies suggest the persistence of symptoms in a large proportion of patients with childhood GERD. However, they were limited by the small sample size and the short duration of follow-up evaluation. We recently completed a cohort study of GERD in 80 individuals diagnosed with childhood GERD during , who after a 10-year follow-up period reported symptoms of GERD and the use of antisecretory medications in a large proportion of patients. 8 That preliminary study was limited by the relatively small sample size, the use of several methods of data collection, the absence of an internal control group, and the low participation rate (39%). Therefore, we conducted a larger study in a different, and more recently identified, cohort, added an internal control group, and used several strategies to increase participation, as well as a uniform method of data collection and several questions about potential GERD risk factors including lifestyle factors, medications, family history, and body mass index. Methods This was a retrospective nested case-control study. The Institutional Review Board for Human Subjects at Baylor College of Medicine reviewed and approved the study. To identify potential participants, we searched the administrative and endoscopic databases at Texas Children s Hospital in Houston, Texas. The administrative database of Texas Children s Hospital contains medical diagnoses recorded in inpatient and outpatient encounters; these diagnoses are coded according to the 9th revision of the International Classification of Disease, Clinical Modification. We identified children with GERD, as defined by erosive esophagitis (530.1), who underwent an upper endoscopic procedure (CPT-4 codes: 43234, 43235, and 43239) in Abbreviations used in this paper: BMI, body mass index; CPT, Current procedural terminology; GERD, gastroesophageal reflux disease; H 2 RA, histimine 2 receptor antagonist; OCP, oral contraceptive pill; PPI, proton pump inhibitor by the AGA Institute /07/$32.00 doi: /j.cgh

2 June 2007 CHILDHOOD GERD 697 Table 1. The Prevalence of GERD Symptoms Among 113 Participants With History of Childhood GERD Number (prevalence per 100) 95% Confidence interval At least weekly heartburn Any heartburn with a duration year At least weekly acid regurgitation At least weekly heartburn or acid regurgitation At least monthly severe heartburn or acid regurgitation PPI with no GERD We excluded patients with cerebral palsy, mental retardation, tracheoesophageal fistula, congenital esophageal stenosis, solid organ transplant, cancer, or cystic fibrosis. We also excluded patients whose residence was outside the Houston area. None of the potential participants were included in our previous study. 8 We obtained information about the time of GERD diagnosis in childhood from the medical records. We also identified pediatric patients with outpatient encounters at Texas Children s Hospital during in whom neither GERD diagnoses nor gastrointestinal endoscopic procedures were recorded. We applied the same exclusion criteria described earlier for children with GERD. This group served as internal controls. We mailed letters to eligible patients explaining the study purpose and requesting a telephone interview; a stamped postcard offering the opportunity to decline also was enclosed. We contracted CODA, Inc. (Durham, NC) to conduct computerassisted telephone interviews using structured questionnaires on potential esophageal and extraesophageal GERD symptoms, comorbid conditions, medications, lifestyle factors (tobacco smoking, alcohol drinking, coffee drinking), family history of GERD, and weight/height. CODA has conducted internal pretests to evaluate the flow of the questions and question wording. Testing of the computer-assisted telephone interview program also was conducted to ascertain that all questions, skip logic, and edits were programmed correctly. The core questions about GERD symptoms were derived from the previously validated Gastroesophageal Reflux Questionnaire. 9 Trained interviewers conducted the surveys. Approximately 10% of each telephone interviewer s work was monitored by a CODA telephone supervisor using silent monitoring equipment. Up to 10 attempts on separate days were made to contact each potential participant before they were considered unresponsive. A similar procedure was used for the internal control group. Data Analysis The prevalence of participants with at least weekly symptoms of heartburn or acid regurgitation was calculated as a proportion of participants with childhood GERD, as well as the internal control group without childhood GERD. Among participants with childhood GERD, we conducted univariate analyses to compare persons with current GERD symptoms with those without GERD symptoms with regard to demographic features (age, sex, race), family history of GERD, weight, height, the use of antisecretory medications (H 2 RA, proton pump inhibitors [PPIs]), oral contraceptive pills (OCPs), and the use of over-the-counter medications. Chi-square tests were used for dichotomous variables, and an unpaired t test was used for continuous variables. Logistic regression models were constructed to examine potential predictors of weekly heartburn or acid regurgitation among participants with childhood GERD. Multivariable models were constructed to simultaneously examine the effect of variables found to be significant in unadjusted analyses. Odds ratios and their accompanying 95% confidence intervals were calculated using the Wald method. Results A total of 160 potential participants with childhood GERD were identified and reached; of those, 9 were deceased, 5 were excluded because of mental or physical comorbidity, and 33 refused to participate. Thus, we had 113 participants who completed the entire telephone-administered questionnaire (70.6% participation rate). There were no statistically significant differences between participants and nonparticipants with regard to age or sex (data not shown). The mean age of the 113 participants at the time of enrollment was 17.8 years (SD, 2.7 y; range, y). Most (73%) were years old and the rest (27%) were years old. The racial distribution was 57% Caucasian, 6% African American, 2% Hispanic, 10% self-identified as other, and 25% of individuals did not report their racial affiliation. There were slightly more females (57.5%) than males. The majority of participants (88.5%) were single or had never been married. Most were high school graduates (27.4%) or had some college or vocational school education (32.7%). Most were students (49.6%) or employed (37.2%). A total of 22.1% of the participants had a history of GERD in first-degree relatives. The mean age in which GERD was diagnosed in childhood was 10.4 years. We identified and reached 71 children with no childhood GERD, of whom 33 completed the survey (47% participation Table 2. Demographic Features in 113 Participants With History of Childhood GERD Compared Between 2 Groups With and Without Current GERD Symptoms Weekly GERD symptoms (N 52) Rest (N 61) P value Age at current interview, median Age at GERD diagnosis, median Female sex (%) 36 (69.2) 29 (47.5).02 Highest level of education 7th 9th grade 0 2 (3.3%).2 10th 11th grade 15 (28.9%) 20 (32.8%) High school graduate 17 (32.7%) 15 (24.6%) Some college or vocational 15 (28.9%) 22 (36.1%) school College graduate 5 (9.6%) 1 (1.6%) Employment status Employed 19 (36.5%) 23 (37.7%).2 Unemployed 7 (13.5%) 6 (9.8%) Homemaker 0 2 (3.9) Student 24 (42.3%) 32 (52.5%)

3 698 EL SERAG ET AL CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 5, No. 6 rate), who served as internal controls. The mean age of the controls was 19.2 years and 52% were female; which was very similar to the case participants. The prevalence of GERD symptoms in children with childhood GERD (1-year recall period) is shown in Table 1. Frequent symptoms were relatively common, approximately 46% of participants reported at least weekly heartburn or regurgitation and a larger proportion (65.5%) reported at least monthly symptoms. An additional 19 (16.8%) with no weekly symptoms used PPIs. The prevalence rates of GERD symptoms were approximately 2-fold higher in children with childhood GERD than controls without childhood GERD; however, because of the small sample size (of controls), the differences fell short of statistical significance. The prevalence of at least weekly heartburn was 5 of 33 (15.2%; 95% confidence interval, 6.5% 31.6%), at least weekly heartburn or acid regurgitation was 10 of 33 (30.3%; 95% confidence interval, 17.1% 47.7%), and at least monthly severe heartburn or acid regurgitation was 2 of 33 (6.1%; 95% confidence interval, 1.5% 21.2%). Twenty-three (20.3%) people reported visiting a physician for heartburn or acid regurgitation for various reasons including severe symptoms in 18 (78.3%), and the belief that they could have a serious disease in 10 (48.5%). A comparison between persons with at least weekly heartburn and/or acid regurgitation (46%) and the rest of the study cohort (54%) is shown in Tables 2 4. Those with symptoms were slightly older and more likely to be female. There were no differences between the groups in education, employment status, or age of GERD diagnosis in childhood (Table 2). We examined the presence of several upper-gastrointestinal and extraesophageal symptoms (Table 3). Stomach pain more than 6 times during the past year was reported to be more frequent and more severe in patients with GERD symptoms. Persons with at least weekly heartburn or regurgitation had more frequent and more severe dysphagia than those without GERD symptoms. As expected, the use of PPIs was considerably more common in patients with frequent GERD symptoms. Chest pain also was reported more frequently in persons with current GERD symptoms than those without these symptoms. However, there were no differences in reporting odynophagia, asthma, cough, wheezing, or hoarseness. Further, there were no significant differences in the proportions of patients with reported Barrett s esophagus, esophageal dilatations, or surgery between the 2 groups (Table 3). There were no significant differences in the proportions of individuals with a history of tobacco smoking, alcohol drinking, or coffee drinking between the groups with and without current GERD symptoms (Table 4). However, coffee drinkers in the GERD group reported significantly higher number of daily coffee drinks than among those without current symptoms. There were also no differences in the family history of GERD, but there was a slight increase in reported Barrett s esophagus Table 3. Gastrointestinal and Extraesophageal Symptoms in 113 Participants With History of Childhood GERD Compared Between 2 Groups With and Without Current Weekly GERD Symptoms Weekly GERD symptoms (N 52) Rest (N 61) P value Esophageal symptoms Dysphagia 24 (46.2%) 18 (29.5%).07 At least weekly dysphagia 3 (5.8%) 4 (6.6%) Dysphagia severity Mild 6/24 (25%) 11/18 (61.1%).06 Moderate 12/24 (50%) 4/18 (22.2%) Severe 6/24 (25%) 3/18 (16.7%) Dilation of the esophagus 4 (7.7%) 6 (9.8%).9 Odynophagia 15/24 (62.5%) 7/18 (38.9%).1 Esophageal surgery 9 (17.3%) 17 (27.8%).2 GERD therapy in past year Antacid or sucralfate 38 (73.1%) 31 (50.8%).01 Any H 2 RA 17 (32.7%) 16 (26.2%).5 Metoclopromide 5 (9.6%) 1 (1.6%).09 Any PPI 29 (55.8%) 19 (31.2%).008 Other upper-gastrointestinal symptoms Stomachache 6 times in past year 32 (61.5%) 19 (31.1%).02 Primary pain above navel 13 (25%) 11 (18.0%).97 Primary pain below navel 7 (13.5%) 7 (18.9%) Primary pain both above and below navel 22 (42.3%) 19 (31.1%) Extraesophageal symptoms At least weekly chest pain 18 (34.6%) 6 (9.8%).001 Mild 11/33 (33.3%) 13/29 (44.8%).5 Moderate 16/33 (48.5%) 10/29 (34.5%) Severe 6/33 (18.2%) 6/29 (20.7%) Globus sensation 16 (30.8%) 10 (16.4%).07 Frequent cough 22 (42.3%) 26 (42.6%).97 Wheezing 23 (44.2%) 26 (42.6%).8 Hoarseness 21 (40.4%) 19 (31.2%).3 Asthma 15 (28.9%) 38 (37.7%).3

4 June 2007 CHILDHOOD GERD 699 Table 4. The Distribution of Potential Risk Factors in 113 Participants With History of Childhood GERD Compared Between Participants With and Without Current Weekly Heartburn or Acid Regurgitation Weekly GERD symptoms (N 52) Rest (N 61) P value Lifestyle factors Tobacco smoking Ever smoked cigarettes regularly 16 (30.8%) 12 (19.7%).2 Age at which smoking started, mean (SD) 16.3 (2.1) 15.1 (2.6) Mean (SD) packs smoked per day when smoking was heaviest 0.5 (.6) 0.3 (.5) Mean (SD) packs currently smoked per day 0.2 (.4) 0.1 (.4) Drinking coffee Drinking coffee 23 (44.3%) 23 (37.7%).5 Mean number of cups per day (SD) 1.3 (1.0) 0.4 (.6) Alcohol drinking Number of alcoholic drinks per week 1 36 (69.2%) 48 (78.7%) (11.5%) 6 (9.8%) (11.5%) 5 (8.2%) 7 4 (7.7%) 2 (3.3%).6 Family history of GERD or its complications Immediate family with GERD 15 (28.9%) 10 (16.4%).3 Mother 9/15 (64.3%) 5/10 (50%).5 Father 6/15 (42.9%) 4/10 (40%).9 Brother 3/15 (21.4%) 0.2 Sister 3/15 (21.4%) 1/10 (10%).6 Immediate family member diagnosed with Barrett s esophagus 2 (3.9%) 0.04 Medications Aspirin 1 tablet per week 15 (28.9%) 11 (18.3%).2 Nonsteroidal anti-inflammatory drug 1 tablet/wk 19 (36.5%) 13 (21.3%).07 Any OCP in female participants 13/36 (36.1%) 4/29 (13.8%).04 Anthropometric features Mean current height, cm (SD) (12.3) (11.2).05 Mean current weight, kg (SD) 68.2 (27.1) 69.9 (21.2).7 Current BMI 23.6 (6.3) 23.4 (6.3).9 BMI 30 7 (13.7%) 7 (11.5%).7 Mean weight 1 year ago, kg (SD) 72.1 (33.9) 68.8 (23.5).6 Mean weight 5 years ago, kg (SD) 63.6 (30.5) 58.3 (19).3 Mean weight 10 years ago, kg (SD) 46.6 (19.1) 40.6 (14.3).1 Mean change in weight from 10 years to current, % (SD) 21.9 (18.3) 29.9 (16.6).04 among participants with current GERD symptoms. There were no differences in current or past height or weight. There were no differences in the intake of aspirin or nonsteroidal antiinflammatory drugs. Although females appeared to be at higher risk of current GERD symptoms, it was explained mostly by oral contraceptive pill (OCP) use. OCP intake was reported more frequently in females with current GERD symptoms than among females without these symptoms. In logistic regression analyses, OCP use was associated independently with the presence of GERD symptoms, adjusting for PPI treatment, age, and these 2 variables together (Table 5). Discussion We report here an average of 8 years of follow-up evaluation for a cohort of 113 patients who were diagnosed with GERD in childhood. We found that GERD in children without underlying severe disorders can persist through adolescence and early adulthood. A large proportion of patients continue to have GERD symptoms and signs, and use antisecretory medications. A total of 65.5% of the study sample complained of monthly symptoms, 46% reported weekly symptoms, and an additional 16.8% with no weekly symptoms were using PPIs. These rates were higher than those obtained in a small sample of controls of a similar age but with no history of childhood GERD. The findings of this study extend and confirm our previous study. 8 However, the current study had several additional advantages: (1) increased participation rate 70.6% as compared with 39%, thus having more generalizable prevalence figures; (2) larger sample size, thus having more accurate prevalence figures; (3) standardization of the questionnaire completion method (all were obtained by telephone whereas the previous study was a mixture of telephone and personal interviews); and (4) inclusion of an internal control group. The considerably higher participation rate likely resulted from identifying patients diagnosed in childhood during , which is a more recent identification period than in our previous study, and also likely resulted from the elimination of the request for an upper endoscopy. The prevalence rates of GERD symptoms in adolescents and young adults in our study population were higher than those

5 700 EL SERAG ET AL CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 5, No. 6 Table 5. The Association Between GERD Symptoms (At Least Weekly Heartburn or Acid Regurgitation) and Female Sex and OCP Use Odds ratio 95% Confidence intervals Female sex (vs male) Unadjusted Adjusted for OCP use OCP use (in 113 participants, of whom 52 have GERD) Unadjusted (vs men) OCP use (in 65 females, of whom 31 have GERD) Unadjusted (vs women on no OCP) Adjusted for age Adjusted for PPI use Adjusted for age and PPI use NOTE. Results are from several logistic regression models. reported by our small group of internal controls (2-fold higher but not statistically significant because of the small number of controls). The prevalence rates reported in this study also are higher than what has been reported previously in a group of that age. Nelson et al 9 conducted a cross-sectional survey of 16 pediatric group practices in the Chicago area and reported the prevalence of heartburn or regurgitation symptoms associated with GER in 3- to 17-year-old children to range between 3.5% and 8.2%. 10 The prevalence rates of symptoms in our study were considerably higher than those reported for children years of age in the study by Nelson et al. 9 Even in the unlikely scenario of all nonrespondents having no symptoms, we estimated that 32.5% of the study population had at least weekly symptoms, which is still higher than the corresponding numbers in the study by Nelson et al. 9 The findings further affirm the relative unimportance of alcohol, coffee, or tobacco smoking in the causation of GERD. Previous studies that linked body mass index (BMI) with GERD did so mostly for obesity (BMI 30), 11 although 1 recent study reported a dose-response relation between BMI and GERD symptoms, even within subjects with normal weight. 12 Although there was no significant association between BMI and GERD symptoms in this study, there were very small proportions of obese individuals in either group and the study was underpowered to detect small differences in BMI. A relatively novel finding is the observed association between oral contraceptive use and GERD symptoms. It seems to explain the higher prevalence of GERD symptoms in females than males. One study obtained sequential measurements of lower esophageal sphincter pressure, basal gastric ph, and fasting plasma gastrin levels in female volunteers who were using oral contraceptives. 13 No changes in basal gastric ph or fasting plasma gastrin levels were observed during any of the periods studied. Lower esophageal sphincter pressure was the same during menses when the volunteers took no medication as during the phase of the cycle when the volunteers were ingesting ethinylestradiol. Lower esophageal sphincter pressure decreased significantly (P.01) during the phase of the cycle when the volunteers took the progestation agent dimethisterone. It therefore was proposed that the progressive increase in plasma progesterone level that occurs during the course of pregnancy might be responsible for the increased incidence of symptomatic heartburn in pregnant women. 13 Lower esophageal sphincter pressure was reported to be reduced at all times during pregnancy, reaching a nadir at 36 weeks and returning to normal during the postpartum period. 14 Despite plausible experimental evidence, there is little information on the epidemiologic association between OCP and GERD. Other studies also have indicated a stronger association between obesity and GERD symptoms among premenopausal women compared with postmenopausal women. 15 This finding needs to be examined further. The findings of this study have to be interpreted within its potential limitations. The current findings may not be generalizable to all children with GERD. The original cohort was identified from patients coded as having erosive esophagitis who were referred to a pediatric gastroenterologist and who underwent an upper endoscopy; therefore, they were likely to have more severe or bothersome symptoms than the rest of the children with GERD. The findings may not be applicable to children with nonerosive esophagitis. Whether the persistence of symptoms through adulthood is associated with progression or permanent changes in the esophageal mucosa is unknown. Given the questionnaire-based nature of the study, we were not able to verify information about medications and body measurements, or assess the severity of GERD by endoscopy or ph testing. Last, our internal control group was small and therefore the differences in GERD prevalence rates were not statistically significant despite a clear trend of higher GERD rates in cases. The recruitment of controls also was more difficult, the participation rate was almost half of that achieved for cases. In conclusion, GERD in normal children persisted into adolescence and early adulthood in a significant proportion of persons, all of whom continued to show GERD symptoms and/or use antisecretory medications. Childhood GERD should be considered a risk factor for adolescent and adult GERD. References 1. Shepherd RW, Wren J, Evans S, et al. Gastroesophageal reflux in children clinical profile, course and outcome with active therapy in 126 cases. Clin Pediatr 1987;26: Treem WR, Davis PM, Hyams JS. Gastroesophageal reflux in the older-child presentation, response to treatment and long-term follow-up. Clin Pediatr 1991;30: Krug E, Bergmeijer JHLJ, Dees J, et al. Gastroesophageal reflux and Barrett s esophagus in adults born with esophageal atresia. Am J Gastroenterol 1999;94: Friedland GW, Sunshine P, Zboralske FF. Hiatal hernia in infants and young children 2-year to 3-year follow-up study. J Pediatr 1975;87: Johnston BT, Carre IJ, Thomas PS, et al. 20-year to 40-year follow-up of infantile hiatal-hernia. Gut 1995;36: Bernhard UA, Shmerling DH. Follow-up examinations of conservatively and surgically treated children with hiatus hernia. Prog Pediatr Surg 1985;18: Ashorn M, Ruuska T, Karikoski R, et al. The natural course of gastroesophageal reflux disease in children. Scand J Gastroenterol 2002;37: El-Serag HB, Gilger M, Carter J, et al. Childhood GERD is a risk factor for GERD in adolescents and young adults. Am J Gastroenterol 2004;99:

6 June 2007 CHILDHOOD GERD Nelson SP, Chen EH, Syniar GM, et al. Prevalence of symptoms of gastroesophageal reflux during childhood a pediatric practicebased survey. Arch Pediatr Adolesc Med 2000;154: Locke GR, Talley NJ, Weaver AL, et al. A new questionnaire for gastroesophageal reflux disease. Mayo Clin Proc 1994;69: Hampel H, Abraham NS, El-Serag HB. Meta-analysis: obesity and the risk for gastroesophageal reflux disease and its complications. Ann Intern Med 2005;143: Jacobson BC, Somers SC, Fuchs CS, et al. Body-mass index and symptoms of gastroesophageal reflux in women. N Engl J Med 2006;354: Van Thiel DH, Gavaler JS, Stremple J. Lower esophageal sphincter pressure in women using sequential oral contraceptives. Gastroenterology 1976;71: Van Thiel DH, Gavaler JS, Joshi SN, et al. Heartburn of pregnancy. Gastroenterology 1977;72: Nilsson M, Johnsen R, Ye W, et al. Obesity and estrogen as risk factors for gastroesophageal reflux symptoms. JAMA 2003;290: Address requests for reprints to: Hashem B. El Serag, MD, MPH, The Houston Veterans Affairs Medical Center, 2002 Holcombe Boulevard (152), Houston, Texas hasheme@bcm.tmc.edu; fax: (713) Dr El Serag is the recipient of a Research Career Development Award from VA Health Services Research and Development (RCD ). The study was funded in part by a grant from TAP Pharmaceuticals and Janssen Eisai (H.B.E.-S.). Dr El Serag also acknowledges support of the Texas Digestive Disease Center (PHS grant DK 56338).