familial clustering of reflux symptoms

ORIGINAL CONTRIBUTIONS

Familial Clustering of Reflux SymptomsNigel J. Trudgill, M.B.Ch.B., M.R.C.P., Kapil C. Kapur, M.B.B.S., M.R.C.P.,and Stuart A. Riley, M.B.Ch.B., F.R.C.P.Department of Gastroenterology, Northern General Hospital, Sheffield, United Kingdom

OBJECTIVE: A number of case reports describe multiple fam-ily members with gastroesophageal reflux disease and Bar-rett’s esophagus. The wider importance of familial factors ingastroesophageal reflux disease has not been established.Therefore, we have studied the prevalence of reflux symp-toms and medication use among relatives of patients withdocumented gastroesophageal reflux disease.

METHODS: A postal questionnaire study of the first degreerelatives of six groups of matched patients. The groupscomprised patients with 1) no dyspeptic symptoms; 2) refluxsymptoms and a normal pH study; 3) reflux symptoms, anabnormal pH study, and a lower esophageal sphincter (LOS)pressure more than 10 mm Hg; 4) reflux symptoms, anabnormal pH study, and a LOS pressure less than 10 mmHg; 5) Barrett’s esophagus; and 6) peptic stricture.

RESULTS: Four hundred eighteen subjects replied (78% re-sponse). Infrequent reflux symptoms were equally commonin all groups of relatives. Frequent reflux symptoms, how-ever, were more common among relatives of patients withan abnormal pH study and normal (26%,p 5 0.007) or lowLOS pressure (27%,p 5 0.01) or Barrett’s esophagus (30%,p 5 0.003), compared with relatives of nondyspeptic pa-tients (9%). Frequent reflux symptoms were no more com-mon among relatives of patients with a normal pH study(16%) or peptic stricture (18%). Reflux medication useshowed a similar pattern.

CONCLUSIONS: Familial clustering of reflux symptoms isseen in relatives of patients with reflux symptoms and in-creased esophageal acid exposure and in relatives of patientswith Barrett’s esophagus. (Am J Gastroenterol 1999;94:1172–1178. © 1999 by Am. Coll. of Gastroenterology)

INTRODUCTION

Despite considerable progress in understanding the patho-physiology of gastroesophageal reflux disease, little isknown of its etiology.

Of potential environmental factors, diet is frequently im-plicated and many patients report symptoms in the post-prandial period. Foods, particularly those with a high fatcontent, cause a decrease in lower esophageal sphincter(LOS) pressure and an increase in the frequency of transientLOS relaxations (TLOSR), thereby facilitating gastroesoph-ageal reflux (1). In a laboratory setting, cigarette smoking

and alcohol consumption may provoke acid reflux in healthyvolunteers and in patients with reflux disease (2–4), butepidemiological studies do not support these associations (5).

Two lines of evidence suggest that familial factors mayplay a role. First, a number of reports detail multiple familymembers with symptomatic or endoscopic reflux disease,Barrett’s esophagus, and esophageal adenocarcinoma (6–13). Second, Romeroet al. (14) have recently reported thatreflux symptoms are more common among relatives ofpatients with Barrett’s esophagus or esophageal adenocar-cinoma, although not among relatives of patients with en-doscopic esophagitis.

Therefore, we studied the prevalence of reflux symptomsin first degree relatives of patients with documented gastro-esophageal reflux, Barrett’s esophagus, and peptic stricture andin symptomatic and asymptomatic matched patient controls.

MATERIALS AND METHODS

SubjectsSix groups of patients were selected. Three groups wereidentified from a database held in the Oesophageal Labora-tory at the Northern General Hospital. All were patients withreflux symptoms who had undergone standardized measure-ment of fasting LOS pressure and ambulatory pH monitor-ing. LOS pressure was determined using a station pull-through technique and derived from the mean of fourmidrespiratory readings from three circumferentially placedsolid state transducers (15). pH monitoring was performedusing an antimony pH probe placed 5 cm above the mano-metrically determined upper border of the LOS. Thirtyconsecutive patients with a total esophageal acid exposureof .5% and a LOS pressure,10 mm Hg (“low” LOSpressure) were selected. Thirty consecutive age- and sex-matched patients with a total acid exposure.5% and a LOSpressure.10 mm Hg (“normal” LOS pressure), and 30 witha total acid exposure,5% and a LOS pressure.10 mm Hgwere then selected.

Thirty consecutive age- and sex-matched patient controls,who had no history of dyspeptic symptoms, were recruitedfrom the Medical Out-Patient Department of the NorthernGeneral Hospital.

Finally, 30 consecutive patients with Barrett’s esophagusand 30 with peptic stricture were identified from endoscopyrecords. Barrett’s esophagus was defined as at least 3 cm

THE AMERICAN JOURNAL OF GASTROENTEROLOGY Vol. 94, No. 5, 1999© 1999 by Am. Coll. of Gastroenterology ISSN 0002-9270/99/$20.00Published by Elsevier Science Inc. PII S0002-9270(99)00116-1

between the proximal margin of the gastric folds and thesquamocolumnar junction at endoscopy, and was confirmedhistologically by finding intestinal metaplasia (16). Pepticstricture was defined as a nonmalignant narrowing associ-ated with histological or endoscopic evidence of esophagi-tis. Patients with Barrett’s esophagus and peptic stricturewere sex matched to the other patient groups.

Each patient was asked to provide the names and ad-dresses of all living first degree relatives over the age of 18yr. Relatives were excluded from the study if their postaladdress was unknown, if they lived abroad, or if they werementally unfit. Relatives who had been pregnant in the pastyear and had had reflux symptoms confined to pregnancywere considered not to have reflux for the purpose of thisstudy. The questionnaire was sent to all identified relatives,with a second mailing 4 weeks later if there was no re-sponse.

The presence of reflux symptoms was defined as eitherheartburn or acid regurgitation during the previous 12months. The use of reflux medication was defined as theconsumption of antacids, H2 antagonists, proton pump in-hibitors, or prokinetics during the previous 12 months.

The effect of social class on reflux symptoms was as-sessed by classifying respondents as manual or nonmanualworkers according to the Standard Occupational Classifica-tion (17). Six groups of drugs were assessed as potentialassociates of reflux symptoms: anticholinergic agents, ben-zodiazepines, calcium antagonists, nitrates, progestagen-containing contraceptives and hormone replacement ther-apy, and theophylline (18).

QuestionnaireA 73-item questionnaire was used. Its form, contents, andvalidation have been described in detail elsewhere (19).Briefly, it measures symptoms of heartburn and acid regur-

gitation during the past year, potential risk factors for refluxdisease, drug therapy, and the demographic details of re-spondents. Minor modifications were made to the question-naire relating to alcohol intake, occupation, age, and gender.

The study was approved by the Ethics Committee of theNorthern General Hospital.

Statistical AnalysisAll data are expressed as median (range) unless otherwisestated. Discontinuous variables were compared using thex2

test and continuous variables using the Mann-Whitney test.Logistic regression analysis was performed to compare theindependent effects of study group, age, gender, body massindex (BMI), alcohol intake, smoking, drug therapy, andoccupational class on the prevalence of reflux symptomsand reflux medication use.

RESULTS

Clinical and Demographic Details ofthe Index Cases, Controls, and RelativesThree patients refused to participate and five had no avail-able relatives. The demographic details of the remaining 172index cases and controls are shown in Table 1.

Five hundred sixty-one relatives were identified and 538fulfilled the study criteria. Four hundred eighteen of the 538(78%) replied, 32 (6%) refused to complete the question-naire. The demographic details of the six groups of respon-dents are shown in Table 2.

Prevalence of Reflux Symptoms andReflux Medication Use Among the RelativesEight respondents who reported reflux symptoms had beenpregnant in the past year and six had had symptoms onlyduring pregnancy.

Table 1. Clinical and Demographic Details of the Index Cases and Controls

No DyspepticSymptoms(n 5 29)

RefluxSymptoms,Normal pH(n 5 29)

RefluxSymptoms,

Abnormal pH,Normal LOS

Pressure(n 5 28)

RefluxSymptoms,

Abnormal pH,Low LOSPressure(n 5 28)

Barrett’sEsophagus(n 5 29)

PepticStricture(n 5 29)

Age (yr) 48 51 49 48 65* 73*Range (21–73) (23–72) (24–77) (23–75) (36–84) (29–88)

Sex (male) 55% 55% 54% 61% 55% 55%Reflux symptoms (%) 0 100 100 100 93 70Endoscopic esophagitis (%) NK 25%† 46%‡ 74% 100% 100%Acid exposure time pH,4 (%) NK 2.3 11.7 14.4 NK NK

(0.1–4.3) (5.2–28.1) (5.6–42.1)LOS pressure (mm Hg) NK 14 13 7 NK NK

(10–28) (10–19) (0–9)Relatives available for study per subject 3 3 3 3 3 2

(1–10) (1–7) (1–7) (1–8) (1–6) (1–6)

* p , 0.001versusother groups.† p , 0.001versusreflux symptoms, abnormal pH, low LOS pressure.‡ p 5 0.05versusreflux symptoms, abnormal pH, low LOS pressure.NK 5 not known.

1173AJG – May, 1999 Familial Clustering of Reflux Symptoms

Infrequent reflux symptoms and use of antacid or pre-scription drug therapy for reflux, at any time during theprevious 12 months, were equally prevalent among therelatives of all six groups (Fig. 1 and 2). However, frequentreflux symptoms or reflux medication use were more com-mon among relatives of patients with reflux symptoms andabnormal pH studies with either normal or low LOS pres-sure, and in the relatives of patients with Barrett’s esopha-gus, than among relatives of patients with no dyspepticsymptoms (Fig. 1 and 2).

Relatives of patients with reflux symptoms and a normalpH study and, surprisingly, relatives of patients with pepticstricture were no more likely to report reflux symptoms orreflux medication use, than the relatives of patients withoutdyspeptic symptoms (Fig. 1 and 2). Logistic regressionanalysis was used to calculate point estimates of the oddsratios (OR) and 95% confidence limits (CI) for reflux symp-toms and reflux medication use among the relatives of thedifferent groups, compared with the relatives of patientswithout dyspeptic symptoms. The results are shown Table 3.

Because the significance of these results may have beeninflated by the lack of independence between members ofthe same family, we repeated the analysis on several subsetsof data, obtained by choosing one individual at random fromeach family. This very conservative method (based on,50% of the available data) confirmed the overall resultsfrom the initial analysis, most notably in respect of relativesof patients with reflux symptoms, an abnormal pH study,and low LOS pressure.

Logistic regression analysis of the study population as awhole revealed that age and BMI were independent predic-tors of both reflux symptoms and reflux medication use. The

Figure 1. Prevalence of reflux symptoms among the groups ofrelatives.

Table 2. Demographic Details of the Relatives

No DyspepticSymptoms(n 5 67)


RefluxSymptoms,


Pressure(n 5 65)

RefluxSymptoms,




Response rate 74% 78% 78% 74% 84% 78%Age (yr) 46 51 40 43 46 52*

Range (20–83) (20–81) (19–82) (18–90) (18–87) (20–90)Sex (male) 39% 40% 45% 38% 39% 42%Manual workers 42% 49% 57% 47% 43% 52%Body mass index 24 25† 24 25 24 26†

Range (17–34) (19–41) (11–50) (17–34) (18–34) (19–45)Excess alcohol intake 4% 12% 9% 10% 15% 4%Current smokers 19% 21% 34% 32% 18% 8%Progestagen-containing drugs 12% 6% 8% 8% 15% 6%Calcium antagonists 4% 4% 5% 8% 3% 2%Nitrates 3% 3% 2% 3% 3% 2%Anticholinergic drugs 3% 4% 2% 3% 0 6%Benzodiazepines 1% 1% 2% 2% 0 0Theophylline 0 1% 0 0 0 0

* p 5 0.04versusno dyspeptic symptoms.† p , 0.01versusno dyspeptic symptoms.Excess alcohol consumption was defined as.21 units per week for a man and 14 for a woman.

1174 Trudgill et al. AJG – Vol. 94, No. 5, 1999

effects of increasing age and BMI on the point estimate ofthe odds ratio of the prevalence of reflux symptoms at leastweekly are shown in Figure 3. Anticholinergic drug therapy

was an independent predictor of reflux symptoms at leastweekly (OR 9.9; 95% CI 2.1–48.1;p 5 0.004). Progestagendrug therapy was also an independent predictor of both atleast daily reflux symptoms (OR 7.2; 95% CI 1.6–33.7;p 50.01) and at least daily reflux medication use (OR 6.4; 95%CI 1.6–25.3;p , 0.01). Gender, smoking, alcohol intake,and occupational class showed no association.

DISCUSSION

Population surveys show that reflux symptoms are highlyprevalent in the community, although most subjects reportinfrequent symptoms (19–23). Therefore, it was not surpris-ing that infrequent reflux symptoms were common in all thegroups of relatives studied, and that differences onlyemerged when more frequent symptoms were analyzed.Relatives of patients with an abnormal pH study with eithera low or normal LOS pressure and patients with Barrett’sesophagus were more likely to report frequent reflux symp-toms. This suggests familial clustering of reflux symptomsamong relatives of patients with objective evidence of acidgastroesophageal reflux or endoscopic evidence of Barrett’sesophagus. There was no evidence of familial clusteringamong relatives of patients with reflux symptoms and anormal pH study. The same pattern emerged when refluxmedication use was analyzed. This clustering was indepen-dent of age, gender, BMI, smoking, alcohol consumption,drug therapy, and occupational class.

Studies such as this are prone to a number of biases (24).To minimize selection bias, consecutive patients were re-cruited from the outpatient clinic and from endoscopy andesophageal laboratory records without knowledge of theirfamily history. Only three (2%) subjects refused to partici-pate. Because families were recruited through an affectedindividual, ascertainment bias may potentially cause an

Figure 2. Prevalence of reflux medication use among the groups ofrelatives.

Table 3. Point Estimate of the Odds Ratio and 95% Confidence Intervals of the Prevalence of Reflux Symptoms and RefluxMedication Use Among the Relatives Compared With the Relatives of the Patients With no Dyspeptic Symptoms


RefluxSymptoms,


Pressure(n 5 65)

RefluxSymptoms,




Any reflux symptoms during past year 1.1 1.6 1.7 1.4 0.7(0.5–2.4) (0.7–3.3) (0.8–3.7) (0.7–2.9) (0.3–1.5)

Reflux symptoms at least monthly 2.0 3.1 5.3 3.6 1.2(0.9–4.8) (1.3–7.4)* (2.2–12.4)† (1.6–8.3)* (0.4–3.1)

Reflux symptoms at least weekly 1.4 4.5 3.9 4.8 1.2(0.5–4.6) (1.5–13.2)* (1.3–11.5)* (1.7–13.4)* (0.4–4.1)

Any reflux medication during past year 1.3 1.2 1.8 1.2 0.6(0.6–2.7) (0.6–2.7) (0.8–3.9) (0.6–2.4) (0.3–1.4)

Reflux medication at least monthly 2.1 4.8 6.2 4.0 1.5(0.7–5.8) (1.7–13.4)* (2.3–16.8)† (1.5–10.5)* (0.5–4.6)

Reflux medication at least weekly 2.6 5.1 6.7 7.4 1.8(0.7–9.7) (1.4–18.9)* (1.9–23.9)* (2.2–25.0)† (0.4–7.2)

* p , 0.01.† p , 0.001.


overestimation of the degree of familial clustering. Theprevalence of reflux symptoms among the relatives of pa-tients without dyspeptic symptoms, however, was similar topublished population surveys (20–23), suggesting the con-trol group was representative of the general population.Finally, as the lack of independence among family membersmay have exaggerated the significance of the results, repeatanalyses were undertaken with one individual randomlychosen from each family. These results confirmed clusteringof reflux symptoms among relatives of patients with in-creased esophageal acid exposure or Barrett’s esophagus.

We were surprised to find that clustering of reflux symp-toms was not seen among the relatives of patients withpeptic stricture, as these patients often have severe refluxdisease and associated dysmotility (25). It might be arguedthat patients with peptic stricture are relatively insensitive toacid reflux and that this insensitivity may be familial (26).However, patients with Barrett’s esophagus are also insen-sitive to acid (27), yet clustering was readily apparent.

A recent study by Romeroet al. (14) has also shownevidence of familial clustering of reflux symptoms amongrelatives of patients with Barrett’s esophagus or esophagealadenocarcinoma, but surprisingly not among relatives ofpatients with endoscopic esophagitis. However, a prelimi-nary report of a community study from the same grouprevealed somewhat contradictory results (5). This studyidentified a higher prevalence of reflux symptoms amongthose who also reported a first degree relative with refluxsymptoms, suggesting that clustering was not confined tofamilies with Barrett’s esophagus and esophageal adenocar-cinoma. Furthermore, several reports describe familieswhose members are affected by Barrett’s esophagus oruncomplicated endoscopic esophagitis (9, 11, 12). We be-lieve the high prevalence of reflux symptoms among thecontrol subjects in the study reported by Romeroet al. (14)may have masked the clustering effect in relatives of pa-tients with reflux esophagitis.

Familial clustering may indicate either shared exposure toan environmental agent or a genetic predisposition. Thepresent study assessed the potential importance of severalenvironmental factors. Smoking and alcohol consumptionprovoke acid reflux and impair acid clearance from theesophagus in a laboratory setting (2–4), but did not emergeas independent risk factors for reflux symptoms and there-fore, are unlikely to account for familial clustering. Al-though subjects with a higher BMI were more likely toreport reflux symptoms, clustering was independent of BMI.Specific dietary factors, such as chocolate and mint, arethought to provoke reflux by direct action on the LOS (28,29). Such dietary factors may potentially impair esophagealfunction, without affecting BMI. Because family membersmay follow similar diets, we cannot exclude dietary factorsas a potential cause of familial clustering.

A genetic predisposition to gastroesophageal reflux dis-ease is also plausible. Several reports have noted familieswhere multiple members have reflux disease, Barrett’sesophagus, or esophageal adenocarcinoma in Barrett’sesophagus (6–13). Four reports suggest an autosomal dom-inant inheritance (9–12). Furthermore, Crabbet al. (9) re-ported several family members with LOS hypotension andesophageal aperistalsis in a family affected by Barrett’sesophagus and reflux disease. It was suggested that aninherited defect in esophageal motility predisposed to gas-troesophageal reflux and its complications. Because manypatients with reflux disease have low basal LOS pressureand impaired esophageal peristalsis, it is tempting to spec-ulate that these motor defects may be genetically deter-mined.

Although the primary aim of our study was to look forevidence of familial clustering, logistic regression analysisof the study population as a whole revealed that age was anindependent predictor of frequent reflux symptoms. A pre-vious community survey has also suggested that refluxsymptoms are more common in the elderly (23), but the

Figure 3. Effects of age and body mass index on the point estimate of reflux symptoms at least once a week in the study population.


mechanisms underlying this association are unclear. LOSpressure does not decrease with age (30) and althoughchanges in peristaltic amplitude (30) and salivary bufferingcapacity (31) have been demonstrated, these changes are ofdoubtful clinical significance. The frequency of sliding hi-atus hernia increases with age (32). Because hiatus herniaimpairs the diaphragmatic component of the LOS and slowsthe clearance of acid from the distal esophagus (33,34), theincreased prevalence may contribute to this age effect.

Increasing BMI was also an independent predictor ofreflux symptoms. It has been suggested that obesity, liketight fitting clothes, may augment the gastroesophagealpressure gradient, thereby encouraging reflux during LOSrelaxation. Studies of weight reduction in patients withreflux and morbid obesity, however, have shown no im-provement in reflux symptoms or acid exposure (35). Webelieve that the higher dietary fat intake of an obese indi-vidual plays a more important role.

In conclusion, in a questionnaire survey of first degreerelatives of patients with gastroesophageal reflux diseaseand matched controls, we have found evidence of familialclustering of frequent reflux symptoms among the relativesof patients with reflux symptoms and increased esophagealacid exposure and among the relatives of patients withBarrett’s esophagus. This suggests either shared exposure toenvironmental agents or a genetic predisposition.

ACKNOWLEDGMENT

We gratefully acknowledge the Mayo clinic for allowing usto use and adapt their questionnaire and the Statistical Ser-vices Unit at the University of Sheffield. This work wassupported by a grant from the Astra Foundation.

Reprint requests and correspondence:Dr. S.A. Riley, NorthernGeneral Hospital, Herries Road, Sheffield, S5 7 AU, UK.

Received June 19, 1998; accepted Nov. 4, 1998.

REFERENCES

1. Dent J, Holloway RH, Tooli J, et al. Mechanisms of loweroesophageal sphincter incompetence in patients with symp-tomatic gastro-oesophageal reflux. Gut 1988;29:1020–8.

2. Kahrilas PJ, Gupta RR. Mechanisms of acid reflux associatedwith cigarette smoking. Gut 1990;31:4–10.

3. Kaufman SE, Kaye MD. Induction of gastro-oesophageal re-flux by alcohol. Gut 1978;19:336–8.

4. Stanciu C, Bennett JR. Smoking and gastro-oesophageal re-flux. BMJ 1972;3:793–5.

5. Locke GR, Talley NJ, Fett SL, et al. Risk factors for gastro-esophageal reflux disease in the community. Gastroenterology1995;108:A25.

6. Everhart CW, Holtzapple PG, Humphries TJ. Barrett’sesophagus: Inherited epithelium or inherited reflux? J ClinGastroenterol 1983;5:357–60.

7. Gelfand MD. Barrett esophagus in sexagenarian identicaltwins. J Clin Gastroenterol 1983;5:251–3.

8. Prior A, Whorwell PJ. Familial Barrett’s oesophagus? Hepato-gastroenterol 1986;33:86–7.

9. Crabb DW, Berk MA, Hall TR, et al. Familial gastroesopha-geal reflux and development of Barrett’s esophagus. AnnIntern Med 1985;103:52–4.

10. Eng C, Spechler SJ, Ruben R, et al. Familial Barrett esophagusand adenocarcinoma of the gastroesophageal junction. CancerEpidemiol Biomarkers Prev 1993;2:397–9.

11. Fahmy N, King JF. Barrett’s esophagus: An acquired condi-tion with genetic predisposition. Am J Gastroenterol 1993;88:1262–5.

12. Jochem VJ, Fuerst PA, Fromkes JJ. Familial Barrett’s esoph-agus associated with adenocarcinoma. Gastroenterology 1992;102:1400–2.

13. Poynton AR, Walsh TN, O’Sullivan G, et al. Carcinomaarising in familial Barrett’s esophagus. Am J Gastroenterol1996;91:1855–6.

14. Romero Y, Cameron AJ, Locke GR, et al. Familial aggrega-tion of gastroesophageal reflux in patients with Barrett’sesophagus and esophageal adenocarcinoma. Gastroenterology1997;113:1449–56.

15. Kraus BB, Wu WC, Castell DO. Comparison of lower esoph-ageal sphincter manometrics and gastroesophageal reflux mea-sured by 24-hour pH recording. Am J Gastroenterol 1990;85:692–6.

16. Spechler SJ, Goyal RK. The columnar-lined esophagus, intes-tinal metaplasia, and Norman Barrett. Gastroenterology 1996;110:614–21.

17. Standard Occupational Classification. Office of PopulationCensuses and Surveys, H.M.S.O., London, 1991.

18. Katzka DA, DiMarino AJ. Pathophysiology of gastroesopha-geal reflux disease: LES incompetence and esophageal clear-ance. In: Castell DO, ed. The esophagus. Boston: LittleBrown, 1995:443–54.

19. Locke GR, Talley NJ, Fett SL, et al. Prevalence and clinicalspectrum of gastroesophageal reflux: A population-basedstudy in Olmsted County, Minnesota. Gastroenterology 1997;112:1448–56.

20. Nebel OT, Fornes MF, Castell DO. Symptomatic gastroesoph-ageal reflux: Incidence and precipitating factors. Am J Dig Dis1976;21:953–6.

21. Thompson WG, Heaton KW. Heartburn and globus in appar-ently healthy people. Can Med Assoc J 1982;126:46–8.

22. Talley NJ, Zinsmeister AR, Schleck CD, et al. Dyspepsia anddyspepsia subgroups: A population-based study. Gastroenter-ology 1992;102:1259–68.

23. Isolauri J, Laippala P. Prevalence of symptoms suggestive ofgastro-oesophageal reflux disease in an adult population. AnnMed 1995;27:67–70.

24. Khoury MJ, Beaty TH, Cohen BH. Fundamentals of geneticepidemiology. Oxford: Oxford University Press, 1993.

25. Stein HJ, Eypasch EP, DeMeester TR, et al. Circadian esoph-ageal motor function in patients with gastroesophageal refluxdisease. Surgery 1990;108:769–78.

26. Winwood PJ, Mavrogiannis CC, Smith CL. Reduced sensitiv-ity to intra-oesophageal acid in patients with reflux-inducedstrictures. Scand J Gastroenterol 1993;28:109–12.

27. Johnson DA, Winters C, Spurling TJ, et al. Esophageal acidsensitivity in Barrett’s esophagus. J Clin Gastroenterol 1987;9:23–7.

28. Murphy DW, Castell DO. Chocolate and heartburn: Evidenceof increased esophageal acid exposure after chocolate inges-tion. Am J Gastroenterol 1988;83:633–6.

29. Sigmund CJ, McNally EF. The action of a carminative on thelower esophageal sphincter. Gastroenterology 1969;56:13–8.

30. Hollis JB, Castell DO. Esophageal function in elderly men—A


new look at “presbyesophagus.” Ann Intern Med 1974;80:371–4.

31. Sonnenberg A, Steinkamp U, Weise A, et al. Salivary secre-tion in reflux esophagitis. Gastroenterology 1982;83:889–95.

32. Stilson WL, Sanders I, Gardiner GA, et al. Hiatal hernia andgastroesophageal reflux. Radiology 1969;93:1323–7.

33. Sloan S, Rademaker AW, Kahrilas PJ. Determinants of gas-troesophageal junction incompetence: Hiatal hernia, lower

esophageal sphincter, or both? Ann Intern Med 1992;117:977–82.

34. Mittal RK, Lange RC, McCallum RW. Identification andmechanism of delayed esophageal acid clearance in subjectswith hiatus hernia. Gastroenterology 1987;92:130–5.

35. Kjellin A, Ramel S, Rossner S, et al. Gastroesophageal refluxin obese patients is not reduced by weight reduction. Scand JGastroenterol 1996;31:1047–51.


familial clustering of reflux symptoms

Documents