Effects of antacid formulation on postprandial oesophageal acidity in patients with a history of episodic heartburn

Transcription

1 Aliment Pharmacol Ther 2002; 16: 435±443. Effects of antacid formulation on postprandial oesophageal acidity in patients with a history of episodic heartburn M. ROBINSON*, S. RODRIGUEZ-STANLEY*, P. B. MINER*, A. J. MCGUIRE, K. FUNG &A.A.CIOCIOLA *Oklahoma Foundation for Digestive Research, University of Oklahoma Health Sciences Center, Oklahoma City, OK, USA; P zer Consumer Healthcare, Morris Plains, NJ, USA Accepted for publication 28 September 2001 SUMMARY Background: Heartburn self-treatment with antacids is extremely common. If the oesophagus is the primary site of antacid action, chewable antacids might raise the oesophageal ph more effectively than swallowable tablets. Aim: To establish a model to assess postprandial acid re ux and to compare the onset and duration of action on oesophageal ph of different antacid formulations. Methods: Twenty subjects with a history of episodic heartburn underwent eight ph monitoring sessions each for 5.5 h postprandially. One hour after consuming a meal consisting of chilli, cheese, raw onions and cola, subjects received 750 mg, 1500 mg and 3000 mg of either chewable or swallowable CaCO 3 tablets, an effervescent bicarbonate solution or placebo. Oesophageal and gastric ph data were collected. Results: Mean intra-oesophageal ph remained lower than baseline for more than 1 h (ph range 5±5.5) postprandially, indicating re ux of somewhat acidic intragastric contents into the oesophagus. The onset of action on oesophageal ph was similar for all antacids (30±35 min). The duration of action on ph varied: chewable tablets and effervescent bicarbonate had relatively long durations of action (oesophagus, 40± 45 min; stomach, 100±180 min); swallowable tablets had little effect. Conclusions: The meal model used in this study dependably produced acidic gastro-oesophageal re ux. Antacids increased oesophageal ph independent of gastric ph, demonstrating that chewing antacids controls oesophageal acidity more effectively than swallowing antacid tablets. INTRODUCTION Heartburn occurs intermittently in more than 30% of otherwise healthy individuals and is almost always associated with acidic gastro-oesophageal re ux. 1±3 In addition, heartburn and related symptoms can be 1, 4, 5 produced by certain `provocative' foods or by overindulgence in food and drink. 6 Correspondence to: Dr S. Rodriguez-Stanley, GI Laboratory Director, Oklahoma Foundation for Digestive Research, University of Oklahoma Health Sciences Center, 711 Stanton L. Young Boulevard, Suite 624, Oklahoma City, OK 73104, USA. sheila-stanley@ouhsc.edu Antacids are popular `over-the-counter' preparations for the relief of episodic heartburn, and have been assumed to assuage heartburn by neutralizing gastric acid, thus preventing subsequent acidic gastro-oesophageal re ux. 7, 8 With large antacid doses (e.g. 156 mmol), fed subjects demonstrate signi cantly elevated gastric ph. 2 However, more recent studies have indicated that usual doses of antacids are primarily active in the distal oesophageal lumen 9, 10 rather than by neutralization of intragastric contents. If this is indeed the case, chewable antacids should rapidly increase and sustain the elevated intra-oesophageal ph better than swallowable antacids. The use of standardized meals eliminates a signi cant variable in re ux and heartburn studies. In a previous Ó 2002 Blackwell Science Ltd 435

2 436 M. ROBINSON et al. comparative study of three meals (hamburger, chilli and sausage-biscuit), 11 the chilli meal induced more re ux episodes and more severe heartburn over the initial 45 min postprandially, as measured on a 100-point visual analogue scale, compared to the other meals. In order to verify the site(s) of action of antacids and to determine the effects of various antacids on oesophageal and gastric ph, increasing doses of swallowable and chewable calcium carbonate tablets were studied using a re uxogenic chilli meal previously demonstrated to produce consistent effects on intragastric and intraoesophageal ph. 11 METHODS This study had two goals: (i) to validate the consistency of a model designed to assess postprandial acid re ux; and (ii) to use this model to compare the onset and duration of action of acid neutralization of various antacid formulations. Subjects Twenty otherwise healthy subjects, 18 years or older, with a history of episodic heartburn on three or more days per week of at least 2 months' duration, were recruited. Study design The protocol was approved by the independent national Western Institutional Review Board of Olympia, Washington, and all subjects signed informed consents prior to the study. Subjects taking antisecretory or motility drugs within 1 week prior to study entry were excluded. Normal physical examinations and laboratory screening tests were required, including a negative pregnancy test in women of childbearing potential. The study was a single-centre, randomized, crossover, placebo-controlled study with each subject acting as his/her own control. Subjects randomly received the following eight study treatments, with study periods separated by at least 24 h: (a) Calcium carbonate (CaCO 3 ) 750 mg chewable tablets (Tums E-X tablets, GlaxoSmithKline, UK); acid neutralizing capacity (ANC), 16 mmol; (b) CaCO mg chewable tablets as above; ANC, 32 mmol; (c) CaCO mg chewable tablets as above; ANC, 64 mmol; (d) CaCO mg swallowable tablets (Glaxo- SmithKline, UK); ANC, 16 mmol; (e) CaCO mg swallowable tablets as above; ANC, 32 mmol; (f) CaCO mg swallowable tablets as above; ANC, 64 mmol; (g) Na + /K + bicarbonate effervescent solution with 32 mmol ANC (the effervescent bicarbonate solution) as Alka-Seltzer Gold tablets (Bayer Corporation, Consumer Care Division, USA); (h) placebo tablets identical to the swallowable tablets. A description of the study protocol is shown in Figure 1. Each subject remained con ned to the research unit throughout each study day. Subjects consumed breakfast by h. From h to h, only water was permitted. Subjects fasted from h until the re uxogenic meal at approximately h. At approximately h, subjects were intubated transnasally with a dual antimony ph electrode catheter (15 cm spacing between electrodes) (Medtronics Synectics, Shoreview, MN, USA), placed to position the proximal electrode 5 cm above the upper margin of the manometrically identi ed lower oesophageal sphincter. All ph data were stored every 4 s for approximately 5.5 h using a portable digital recorder (Mark III Gold, Medtronic Synectics, Shoreview, MN, USA). Electrodes were calibrated to ph 1.07 and 7.01 Figure 1. Diagrammatic example of the protocol for the study.

3 ANTACID FORMULATION AND OESOPHAGEAL ACIDITY 437 with solutions composed of 59 mm KNO 3,27mM KCl (ph 1.07) and 16.5 mm Tris buffer, 40 mm KNO 3, 96 mm KCl (ph 7.01). Stored data were processed in DOS mode using temperature-compensating ph recording software (Polygram for Windows, Version 2.0, Medtronic Synectics, Shoreview, MN, USA). One hour after intubation, at approximately h, a re uxogenic meal was consumed over 30 min. 11 The meal consisted of at least 227 g of Wendy's TM chilli supplemented with half a cup of cheddar cheese and half a cup of chopped raw onions and 340 ml Coke TM Classic cola. Additional portions of chilli and cola were allowed (in half portions), but each subject was required to consume the same amount at each subsequent study session. The study medication was administered 1 h postprandially at h. Tablets were chewed, swallowed whole with water or dissolved in water and the liquid swallowed, according to the randomization treatment code. Three hours after administration of the study medication (at h), ph recording was terminated and the ph electrodes were removed. During the 5-h observation period, subjects did not recline at an angle of more than 30 from upright. Untoward events were monitored for safety (none occurred). Study analyses Intragastric ph. Intragastric ph was recorded every 4 s. All data points between 0 and 5 min (75 values) were used to calculate the mean ph for this time point. For each treatment and each subject, 5-min average oesophageal and gastric ph values were calculated and the 5-min ph averages were computed over the entire study period. These mean ph values were then averaged for all subjects at the speci ed time points (every 5 min) and over the speci c time periods as described below. For each subject, the difference was calculated between 5-min averages for each active treatment and placebo. These differences were tested for normality using the Shapiro±Wilks procedure. 12 The differences were then analysed using a paired t-test if the normality assumption was not rejected, or a Wilcoxon signed-rank test if the normality assumption was rejected. The onset of action, duration of action and mean intragastric ph over speci c time periods were all computed as described below. Onset and duration of action. Analyses were carried out separately for intragastric and intra-oesophageal ph, producing estimates of the onset and duration of action for each active treatment in the stomach and in the oesophagus. The onset of action is the time from dosing to the rst statistically signi cant difference vs. placebo (P < 0.05). The offset time is the time to the rst nonsigni cant difference following the onset time. The duration of action is de ned as the offset time minus the onset time. Speci c time periods before and after dosing. For each pre-treatment and post-treatment time interval, oesophageal and gastric ph averages were calculated. Mean pre-treatment ph was calculated for two time periods: 5±60 min after intubation (i.e. )145 to )90 min) and the 60 min before administration of the study medication (i.e. )60 to 0 min). Mean ph was calculated for four post-treatment time periods: 0±30, 0±60, 0±90 and 0±120 min post-dosing. Each mean ph was calculated for intra-oesophageal and intragastric ph. Mean ph values were compared using univariate analysis of variance (ANOVA) with the factors sequence, period and treatment. All pairwise treatment comparisons were then made using Fisher's least signi cant difference procedure. RESULTS Demographics Twenty subjects (11 males, nine females) completed all eight treatments. The mean age was 44 years, with a range of 24±63 years. At the time of study initiation, 15 of 20 subjects were using antacids for the treatment of heartburn. Ten were taking calcium-rich Rolaids and ve were using the antacid Mylanta. These pre-study medications were deemed to have no signi cant effect on subsequent study results and all such medications were discontinued for the duration of the study period. Pharmacodynamic data: intra-oesophageal and intragastric ph pro les The mean 5-min oesophageal and gastric ph values over the 5.5-h study period are presented in Figures 2 and 3, respectively. Statistical analyses showed that there were no signi cant differences within dosage forms (i.e. various chewable doses vs. the swallowable

4 438 M. ROBINSON et al. Figure 2. Plot of mean intra-oesophageal ph/5 min during the entire recording period. Subjects consumed the re uxogenic meal from 60 to 90 min after the start of the study (shown by the shaded area). Antacids or placebo were given at 150 min after the start of the study (Dose). Data for swallowable tablets and for chewable tablets have been combined. Figure 3. Plot of mean intragastric ph/ 5 min during the entire recording period. Subjects consumed the re uxogenic meal from 60 to 90 min after the start of the study (shown by the shaded area). Antacids or placebo were given at 150 min after the start of the study (Dose). Data for swallowable and for chewable tablets have been combined. dosage levels). Therefore, the data for chewable tablets and for swallowable tablets have been combined for display in these gures. The 5-min oesophageal and gastric ph averages for 1 h after dosing are presented for each dosage form in greater detail in Figures 4 and 5, respectively. Overall, effervescent bicarbonate rapidly elevated oesophageal ph compared with the other active treatments, and chewable CaCO 3 formulations sustained ph elevation vs. swallowable CaCO 3 (Figures 2 and 4; Table 1). Chewable calcium carbonate tablets and the effervescent antacid solution more effectively raised oesophageal ph than any of the swallowable formulations. In the stomach, only the effervescent bicarbonate and higher doses of chewable CaCO 3 appeared to reliably elevate ph compared with the other antacid treatments and placebo (Figures 3 and 5; Table 1). Figures 4 and 5 show the 5-min means for oesophageal and gastric ph during the rst hour after dosing. Figure 4 shows that, within 5 min of dosing, the effervescent solution and chewable calcium carbonate tablets produced a rise in ph in the oesophagus to above 6.5 that was sustained for up to 60 min. Swallowable calcium carbonate tablets produced a slower and less marked rise in ph that was not sustained over this period. Figure 5 shows a marked increase in intragastric ph only with the effervescent bicarbonate solution. There were less marked and inconsistent increases in intragastric ph with the swallowable and chewable forms of calcium carbonate.

5 ANTACID FORMULATION AND OESOPHAGEAL ACIDITY 439 Figure 4. Mean intra-oesophageal ph during the rst hour after dosing (at 150 min). Symbols for each individual treatment are shown in the diagram; Chew, chewable tablets; Swallow, swallowable tablets; Effervescent, effervescent solution. Figure 5. Mean intragastric ph during the rst hour after dosing (at 150 min). Symbols for each individual treatment are shown in the diagram; Chew, chewable tablets; Swallow, swallowable tablets; Effervescent, effervescent solution. Table 1. Mean intra-oesophageal and intragastric ph with calcium carbonate formulations, effervescent bicarbonate solution or placebo for the period 0±60 min after dosing (n ˆ 20) Chew Swallow Placebo Effervescent 750 mg 1500 mg 3000 mg 750 mg 1500 mg 3000 mg Oesophageal ph * 5.71* 5.76* 5.89* 5.42* Gastric ph * * 3.01* 3.01* * *P < 0.05 vs. placebo. P < 0.05 vs. swallowable 750, 1500 mg. Effects of the re uxogenic meal A previous study demonstrated that a chilli meal reliably produces re ux and heartburn in this population. 11 The present study veri ed the re uxogenic nature of a chilli meal. The analysis of intra-oesophageal and intragastric ph after meals is often dif cult due to the inherent variability of ph: the ingested food buffers and dilutes gastric acid against a background of rapidly rising stimulated acidity as the stomach empties. However, the following generalities can be made regarding the effects of this meal on intra-oesophageal and intragastric ph. Before test meal ingestion, the intra-oesophageal ph remained stable with a value of approximately 6.7 (mean values for all treatment groups were in the range 6.68±6.88). This fell to a value of approximately ph 5 during meal ingestion (see Figures 2 and 4, 0±60 min).

6 440 M. ROBINSON et al. The actual ph of the test meal was 5.5 (measured by homogenizing the constituents). The oesophageal ph remained low compared to baseline (mean values for the treatment groups were in the range 5.21±5.40) for the 60-min period following consumption of the test meal, indicating the re ux of gastric contents and substantiating the ability of the test meal to provoke acidic re ux. As expected, gastric ph was low before the test meal (mean ph for the groups was in the range 1.49±1.78) and rose to approximately ph 5 during the meal. Postprandially, this was followed by a gradual fall towards pre-meal ph values that did not reach pre-meal values before antacid dosing (see Figures 3 and 5). During the 60-min period after test meal consumption, and before antacid dosing, mean ph values for the treatment groups ranged from 3.98 to Onset and duration of action The onset of action and duration of action were estimated separately for oesophageal and gastric ph. Calculated times to onset of a signi cant reduction in oesophageal and gastric acidity were similar for all doses and product forms, except that the low doses (750 mg) of chewable and swallowable antacid formulations had minimal effects on gastric ph. Swallowable calcium carbonate 750 mg had no measurable onset time for gastric ph. The onset of a signi cant reduction in oesophageal and gastric acidity (increase in ph) for all other active treatments was between 30 and 35 min. The duration of action varied widely between products (Figure 6). In general, chewable CaCO 3 forms had a substantially longer duration of action (up to four times longer) than the same dose of swallowable CaCO 3 for the neutralization of oesophageal acid. For example, chewable CaCO mg and effervescent Na + /K + bicarbonate had relatively persistent effects (ranges: 40 min in the oesophagus and 100±180 min in the stomach), while swallowable 750 mg and 1500 mg CaCO 3 only transiently altered ph (ranges: 10±15 min in the oesophagus and 5±35 min in the stomach). Speci c time intervals for post-treatment ph For the period 0±30 min after dosing, the mean intraoesophageal ph ranged from 4.88 to From Figure 4, the maximum mean ph appeared to be highest for the effervescent solution at around ph 7; for the chewable tablets, this value was ph 6.5 and, for the swallowable tablets, it was ph 6±6.5. The mean intragastric ph ranged from 2.85 to Effervescent solutions produced the highest rise in mean intragastric ph (ph 4.7; Figure 5). Statistical analyses of ph for this period did not show any signi cant differences in mean oesophageal or gastric values. The analysis of the three post-treatment periods (0±60, 0±90 and 0±120 min) produced similar results, and only those for the period 0±60 min are presented in Table 1. During the rst 60 min after treatment, the following treatments showed statistically signi cantly greater mean oesophageal ph (Table 1; P < 0.05) compared to placebo: chewable 750 mg, 1500 mg and 3000 mg, swallowable 750 mg and 3000 mg and effervescent bicarbonate solution. Effervescent bicarbonate solution and chewable 750 mg, 1500 mg and 3000 mg showed statistically signi cantly greater Figure 6. Summary of duration of action (minutes) for calcium carbonate formulations and effervescent bicarbonate on oesophageal and gastric ph. n ˆ 20. The duration of action for chewable 3000 mg continued until the end of the observation period. Ch, chewable antacid tablets; Sw, swallowable antacid tablets; Bicarb, effervescent bicarbonate solution.

7 ANTACID FORMULATION AND OESOPHAGEAL ACIDITY 441 mean ph values than swallowable 750 mg and 1500 mg (Table 1; P < 0.05). During this period, there were also signi cant differences among the mean gastric ph values of the eight treatments, with chewable 3000 mg, swallowable 750 mg and 3000 mg and the effervescent bicarbonate solution all neutralizing acid signi cantly better than placebo (Table 1). The effervescent bicarbonate solution achieved signi cantly greater elevation of intragastric ph than the other six active treatments (P < 0.05). DISCUSSION Numerous studies, 4, 7±10 including the present one, have demonstrated that antacids probably relieve heartburn pain by neutralizing acid directly within the oesophageal lumen and not necessarily within the gastric lumen. It is reasonable to suggest, therefore, that even small doses of chewed antacid tablet should provide more effective oesophageal ph control than quite large doses of swallowed antacids. The present study supports this suggestion and also explains the otherwise counterintuitive notion that antacids with lower ANC might surpass those with far higher ANC. In our study, the onset of antacid effects seemed far slower than might be anticipated. It is likely, however, that the presence of food in the stomach with re uxed food in the oesophagus (from the early postprandial period) undoubtedly obscures, to some extent, the true antacid onset. Antacids are widely used for the self-treatment of episodic heartburn, 13±15 although many physicians tend to dismiss the therapeutic value of such `low tech' therapy. Most over-the-counter antacids used by consumers range between 15 and 40 mmol ANC per dose. The doses of antacids selected for this study represent a range of ANCs used by consumers. For example, Tums, Tums E-X and Tums Ultra contain calcium carbonate with total ANC values of 10, 15 and 20 mmol per tablet. 16 Rolaids antacid tablets contain calcium carbonate and magnesium hydroxide with an ANC of 14.7 mmol per tablet. 16 Maalox and Maalox extra-strength tablets contain aluminium and magnesium hydroxide with ANCs of 10 and 23 mmol per tablet, respectively. 16 Mylanta in a gelcap formulation contains calcium carbonate and magnesium hydroxide with an ANC of 23 mmol when taking two tablets. 16 These doses are far lower than those employed in the classical studies by Fordtran et al. for duodenal ulcer disease treatment. 7 Individuals participating in this study had a history of food- and beverage-related episodic heartburn and tended to self-medicate with over-the-counter products. Such frequent users of antacids in the community have previously been characterized 15, 17 and, despite clear evidence of acidic gastro-oesophageal re ux as the basis of such antacid use, more than half of such individuals have no evidence of even trivial erosive oesophagitis. 17 The use of a standardized and validated provocative meal in the present study facilitated a comparative de nition of the onset, duration and extent of antacid activity. Many studies have used provocative meals with various noxious components to evaluate antacids and antisecretory drugs for heartburn prophylaxis and/or treatment. 9, 10, 18, 19 The lack of a standardized meal has led to studies of the effects of several meals designed to reliably produce gastro-oesophageal re ux and heartburn. 11 The re uxogenic meal used in this study (chilli with cheese and onions, and a caffeinated soft drink) was slightly modi ed from the report by Rodriguez et al., 11 and contained no measurable capsaicin which could be considered a noxious component. In this previous study, all meals produced re ux and heartburn postprandially. However, the chilli meal induced more severe heartburn compared to baseline vs. the other meals, at 15±45 min postprandially. Chilli also induced more re ux episodes and resulted in oesophageal ph < 4 for 17% of the postprandial time period. Because the chilli meal was previously shown to produce symptoms, symptoms were not recorded in the present study. In the current study, mean oesophageal ph dropped rapidly following the meal, and remained at 5±5.5 for more than 60 min after eating, indicating re ux of meal contents (meal ph is approximately 5.0). As expected, intragastric ph pro les were low prior to the meal and rose to approximately ph 5 (the ph of the meal), followed by a gradual fall to pre-meal ph values prior to antacid dosing. The results indicate that effervescent sodium bicarbonate solution and chewable antacids provide signi cant elevation of intra-oesophageal ph, while swallowable calcium carbonate leads to minimal or no intra-oesophageal ph elevation. Of the antacid formulations studied, only effervescent Na + /K + bicarbonate solution reliably elevated intragastric ph. Neither swallowable nor chewable antacid tablets led to statistically signi cant increases in intragastric ph, and the actual rise in intragastric ph with swallowable or

8 442 M. ROBINSON et al. chewable antacids was minimal over time, as demonstrated in Figure 3. These data provide further evidence that antacids probably affect acid re ux and related heartburn symptoms directly within the oesophageal lumen, at least in part by neutralization of re uxed acid, rather than by any impact on intragastric acidity. In this study, calcium carbonate tablets manifested little discernible dose±response effect on oesophageal or gastric ph in doses ranging from 750 to 3000 mg. The only dose-dependent activity apparent in this study was an increase in the duration of action with higher doses of chewable CaCO 3. This suggests that 750 mg of chewable calcium carbonate may be the optimal dose and form to control intra-oesophageal ph in subjects with intermittent or episodic heartburn or acid re ux. 9, 10 Several earlier studies compared the effects of different antacids on intragastric and intra-oesophageal ph using various provocative meals. Chewable tablets containing either aluminium/magnesium hydroxide or calcium carbonate and two liquid antacids containing aluminium/magnesium hydroxide differing in ANC were studied. Similar results were demonstrated in both of these studies: a rapid and prolonged increase in oesophageal ph coupled with modest changes in intragastric ph. As with the current study, these data strongly suggest that the lower oesophagus is the primary site of antacid action for heartburn relief. Although there have been no de nitive individual studies reporting a direct correlation between the control of intragastric ph/intra-oesophageal ph and symptom relief, meta-analyses by the group at McMaster University in Ontario 20, 21 suggest a strong relationship between the level of acid suppression or neutralization and symptom relief as well as mucosal healing (when applicable). In a study recently published by our group, we found that there was a close temporal relationship between heartburn and oesophageal acidity, but that a relatively large decrease in oesophageal acid exposure appeared to be necessary to produce a statistically signi cant decrease in heartburn severity. 22 The present study was conducted in a pertinent target population, i.e. one with a history of frequent heartburn using over-the-counter antacids for relief. This group comprises a very large segment of heartburn sufferers in the community. Many such individuals have no discernible mucosal damage if endoscopy is undertaken. 23 It may be concluded that chewable and effervescent antacids lead to rapid elevation of intra-oesophageal ph and that such effects may or may not be associated with concomitant changes in gastric ph. The review of ph records showed a rapid rise of oesophageal ph within 5 min of antacid administration, although this did not become statistically signi cant until later postprandially. It is not known whether a larger population might have shown an earlier onset of effects or if the effects of the meal and saliva would always obscure early antacid effects. The admixture of saliva and chewable antacid would be expected to be more effective than even quite large doses of swallowable antacid tablets. It should be possible to develop improved antacid formulations by increasing oesophageal mucoadherence, if this is indeed one mode of action of antacids. ACKNOWLEDGEMENTS This study was supported by P zer Consumer Healthcare, NJ, USA and GlaxoSmithKline, NC, USA. REFERENCES 1 Nebel TO, Castell D. Lower esophageal sphincter pressure changes after food ingestion. Gastroenterology 1972; 63: 778±83. 2 Dent J, Dodds WJ, Friedman RH, et al. Mechanism of gastroesophageal re ux in recumbent asymptomatic human subjects. J Clin Invest 1980; 65: 256±67. 3 Dodds WJ, Dent J, Hogan WJ, et al. Mechanism of gastroesophageal re ux in subjects with re ux esophagitis. N Engl J Med 1982; 301: 1547±52. 4 Lanza FL, Smith V, Page-Castell JA, Castell DO. Effectiveness of foaming antacid in relieving induced heartburn. South Med J 1986; 79: 327±30. 5 Saunders JHB, Oliver RJ, Higson DI. Dyspepsia: incidence of non-ulcer disease in a controlled trial of ranitidine in general practice. Br Med J 1986; 292: 665±8. 6 Koelz HR, Blum AL. Swiss Esophagitis Study Group. Cardinal symptoms of re ux esophagitis. Gastroenterology 1986; 90: Fordtran JS, Morawski SG, Richardson CT. In vivo and in vitro evaluation of liquid antacids. N Engl J Med 1973; 288: 923±8. 8 Graham DY, Patterson DJ. Double-blind comparison of liquid antacid and placebo in the treatment of symptomatic re ux esophagitis. Dig Dis Sci 1983; 28: 559±66. 9 Decktor DL, Robinson M, Maton PN, Lanza FL, Gottlieb S. Effects of aluminum/magnesium hydroxide and calcium carbonate on esophageal and gastric ph in subjects with heartburn. Am J Ther 1995; 2: 546± Decktor DL, Robinson M, Gottlieb S. Comparative effect of liquid antacids on esophageal and gastric ph in subjects with heartburn. Am J Ther 1995; 2: 481±6. 11 Rodriguez S, Miner P, Robinson M, Greenwood B, Maton P, Pappa K. Meal type affects heartburn severity. Dig Dis Sci 1998; 43: 485±90.

9 ANTACID FORMULATION AND OESOPHAGEAL ACIDITY Shapiro SS, Wilks MB. An analysis of variance test for normality (complete samples). Biometrika 1965; 52: 591± Moore EW, Scarlata RW. The determination of gastric acidity by the glass electrode. Gastroenterology 1965; 49: 178± Makhlouf GM, Blum AL, Moore EW. Undissociated acidity of human gastric juice. Gastroenterology 1970; 58: 345± Graham DY, Smith JL, Patterson DJ. Why do apparently healthy people use antacid tablets? Am J Gastroenterol 1983; 78: 257± Physicians' Desk Reference for Non-Prescription Drugs, 18th edn. Montvale NJ: Medical Economics Company, Inc., Robinson M, Earnest D, Rodriguez-Stanley S, et al. Heartburn requiring frequent antacid use may indicate signi cant illness. Arch Intern Med 1998; 158: 2373±6. 18 Pappa KA, Williams BO, Payne JE, Buaron KS, Mussari KL, Ciociola AA. A double-blind, placebo-controlled study of the ef cacy and safety of non-prescription ranitidine 75 mg in the prevention of meal-induced heartburn. Aliment Pharmacol Ther 1999; 13: 467± Gottlieb A, Decktor AL, Eckert JM, Simon TJ, Stauffer L, Ciccone PE. Ef cacy and tolerability of famotidine in preventing heartburn and related symptoms of upper gastrointestinal discomfort. Am J Ther 1995; 2: 314±9. 20 Hunt RH. The relationship between control of ph and healing and symptom relief in gastro-esophageal re ux disease. Aliment Pharmacol Ther 1995; 9(Suppl. 1): 3±7. 21 Howden CW, Burget DW, Hunt RH. Appropriate acid suppression for optimal healing of duodenal ulcer and gastroesophageal re ux disease. Scand J Gastroenterol 1994; 2(Suppl. 201): 79± Robinson M, Rodriguez-Stanley S, Ciociola AA, et al. Synergy between low-dose ranitidine and antacid in decreasing gastric and oesophageal acidity and relieving meal-induced heartburn. Aliment Pharmacol Ther 2001; 15: 1365± Murdock RH, Pappa KA, Giefer EE. Endoscopic ndings in a target population for over-the-counter treatment of heartburn. Gastroenterology 1994; 106(4pt2): A146.