Mechanisms of gastroesophageal reflux in healthy premature infants

M

Mechanisms of gastroesophageal reflux in healthy

premature infants Taher I. Omari, BSc, PhD, Christopher Barnett, MBBS, FRACP, Antonie Snel, RN, RM, Wendy Goldsworthy, RN, RM, Ross Haslam, MBBS, FRACP, Geoff Davidson, MBBS, MD, FRACP, Chellam Kirubakaran, MBBS, MD, Malcolm Bakewell, Robert Fraser, MBBS, PhD, FRACP, and John Dent, MBBChir, PhD, FRACP, FRCP

Objectives: The aim of this study was to characterize the motor events responsible for gastroesophageal reflux (GER) and esophageal acid clearance in a cohort of healthy preterm infants.

Study design: Esophageal motility was recorded for 2 to 3 hours after a feeding in 24 preterm infants, 31 to 38 weeks’ postmenstrual age, by using a sleeveside hole micromanometric assembly incorporating a pH probe.

Results: Sixty acid GER episodes were recorded by pH probe, and 133 nonacid GER episodes were recorded manometrically by the presence of esophageal common cavities. Of the 193 GER episodes, 159 (82%) were associated with transient lower esophageal sphincter relaxation (TLESR). TLESRs were significantly longer in duration than single swallows (15.1 seconds vs 5.6 seconds, P < .001) and had lower nadir pressures (0.8 mm Hg vs 2.3 mm Hg, P < .001). A total of 3216 esophageal body pressure waves were analyzed; 70% of swallow-induced pressure waves were peristaltic in sequence compared with 5% of swallow-unrelated pressure waves. During periods of esophageal acidification (pH below 4), peristaltic esophageal body pressure wave sequences cleared acid refluxate more effectively than nonperistaltic pressure wave sequences. Conclusions: In healthy preterm infants, TLESRs are the predominant mechanism underlying GER, and esophageal clearance mechanisms are well developed by at least 31 weeks’ postmenstrual age. (J Pediatr 1998;133:650-4)

From the Gastroenterology Unit, Neonatal Medicine Unit, and Department of Biomedical Engineering, Women’s and Children’s Hospital, North Adelaide Australia; and Gastrointestinal Medicine, Royal Adelaide Hospital, Adelaide, Australia. Supported by the National Health and Medical Research Council of Australia, Women’s and Children’s Hospital Research Foundation, Queen Victoria Hospital Research Foundation, Channel 7 Children’s Research Foundation of South Australia. Submitted for publication Feb 10, 1998; revision received July 13, 1998; accepted Aug 28, 1998. Reprint requests: Taher I. Omari, BSc, PhD, Gastroenterology Unit, Women’s and Children’s Hospital, North Adelaide, Australia, 5006. Copyright © 1998 by Mosby, Inc. 0022-3476/98/$5.00 + 0 9/21/94145

650

Gastroesophageal reflux disease is common in premature infants and is associated with irritability, frequent vomiting, apnea, aspiration pneumonia, and failure to thrive.1 Gastroesophageal reflux may also exacerbate the effects of bronchopulmonary dysplasia and has been implicated in the pathophysiology of sudden infant death syndrome.1 Recent studies in healthy preterm infants have shown that the tonic pressure at the gastroesophageal junction generated by the lower esophageal sphincter is sufficient to maintain esophagogastric competence.2-4 In adults and older chil-

dren, simultaneous esophageal manometry and pH monitoring have shown that GER is usually due to transient LES relaxation; whereas other mechanisms, including reduced basal sphincter tone, account for a minority of reflux episodes.5-11 To date, there have been no similar data in premature infants because of the lack of appropriate techniques for recording the motor events during GER episodes in this age group. GER LES LESP LESR PMA TLESR

Gastroesophageal reflux Lower esophageal sphincter Lower esophageal sphincter pressure Lower esophageal sphincter relaxation Postmenstrual age Transient lower esophageal sphincter relaxation

Our laboratory has recently developed perfused micromanometric assemblies, incorporating sleeve sensors, which are suitable for studies of premature infants.2,3,11,12 These assemblies incorporate a single larger lumen that can be used for gavage feeding. Initial data with this methodology indicate that the regulatory mechanisms responsible for maintenance of LES resting tone and relaxation are developed in infants as young as 33 weeks’ gestation.2,3 The aim of this study was to use combined manometry and pH recording to characterize the motor mechanisms responsible for acid GER and esophageal acid clearance in healthy premature infants.

METHODS Subjects The study protocol was approved by the Ethics Research Committee of the

THE JOURNAL OF PEDIATRICS VOLUME 133, NUMBER 5 Women’s and Children’s Hospital, and written informed parental consent was obtained before each study. Studies were performed in 24 (13 male and 11 female) healthy preterm infants with a mean postmenstrual age of 35 weeks (range, 31 to 38 weeks). All infants were well at the time of the study, were considered not to have reflux disease, had no evidence of neurologic dysfunction, and were not receiving prokinetic medication. Mean infant weight was 1948 g (range, 1290 to 2990 g). Seven infants were receiving either theophylline or caffeine treatment because of previous episodes of apnea and bradycardia. Thirteen infants were being gavage (tube) fed with unfortified expressed breast milk and 11 with infant formula (Enfalac 20/24 calorie, Mead Johnson, Canada).

OMARI ET AL

Table I. Number of motor events recorded in association with GER

Acid GER (%) GER mechanism

Non-acid GER (%)

TLESR 116 (87.2) Swallow-induced LESR Single swallow 1 (0.8) Multiple swallows 9 (6.8) Peristaltic failure 2 (1.5) Observed vomiting 5 (3.8) Hypotonic LES (<5 mm Hg) — Contracted LES (≥5 mm Hg) — Total 133

software, Royal Adelaide Hospital, C. Malbert).

Protocol Manometric Technique The manometric technique used in this study, previously described and validated, allows monitoring of pressures at very low rates of manometric infusion with an assembly that could also be used to feed the infants.2,3 The 2.0 mm outer diameter intraluminal portion of the manometric assembly consisted of a 10-lumen extrusion, which incorporated a 2.5-cm long sleeve sensor for LES pressure measurement.13 Six side-hole sensors (located at –1.5 cm, 1.5 cm, 3.0 cm, 4.5 cm, 9.5 cm, and 10.5 cm relative to the sleeve midpoint) recorded pressures from the pharynx, esophageal body, and stomach. Pharyngeal side holes were air-perfused at 5 mL/min, and all other side holes were perfused with sterile degassed water at 0.01 or 0.02 mL/min (esophageal) and 0.02 or 0.04 mL/min (sleeve and stomach) with hydraulic resistors (Dentsleeve Pty Ltd, Parkside, South Australia). Bubble entrapment was minimized by preflushing of manometric lines and catheters with carbon dioxide.12 Esophageal pH was monitored with an antimony pH probe (outer diameter, 1.5 mm; Synectics, Stockholm, Sweden), positioned 5 cm proximal to the center of the LES sleeve. Analog pressure transducer and pH probe signals were amplified and filtered with a Synectics polygraph. Data acquisition and analysis were performed on a Macintosh Quadra 700 computer with software based on National Instruments’ Labview (M.A.D.

The assembly was passed through the nose and positioned with the midpoint of the sleeve straddling the LES high-pressure zone. After the assembly was positioned and with the infant in the right lateral position, the feed was administered over 15 to 30 minutes through the core channel of the assembly. Esophageal pH and spontaneous esophageal body and LES motor patterns were then recorded during the feed and then for 2 to 3 hours after the feed (mean recording time, 2.4 hours). The assembly is usually well tolerated by infants, with no adverse effects related to the procedure. Observational notes on infant behavioral patterns were taken, and the occurrence of vomiting or regurgitation was also noted.

Analysis of Manometric Tracings The postprandial manometric recordings were analyzed to assess esophageal and LES motor function, as well as mechanisms of GER and GER clearance.

CHARACTERIZATION OF ESOPHAGEAL PRESSURE WAVES. Esophageal body pressure waves that exceeded a threshold amplitude of 10 mm Hg were identified and analyzed. The propagation of esophageal pressure wave sequences was defined by the timing of the onset of the pressure wave upstroke. Pressure wave sequences were characterized as peristaltic or nonperistaltic (synchronous, retrograde,

No strain

Strain

Total

22

20

43 (74.0)

— 2 3 2 — — 29

— 1 4 3 1 2 31

— 3 (5.0) 7 (11.7) 5 (8.3) 1 (1.7) 2 (3.3) 60

or incomplete) according to criteria already established for premature infants.2,3 Pressure waves were classified as swallowinduced if a swallow was recorded in the 5 seconds before the onset of the most proximal contraction and as swallow-unrelated if they were not associated with swallowing. Pressure wave sequences that occurred within 15 seconds of the previous swallow were not analyzed because of the inhibitory effects of prior peristalsis.

DETERMINATION OF LESP AND CHARACTERIZATION OF LESR. LESP was defined as the difference between end-expiratory LESP and gastric pressure. Swallow-related LESRs were defined as those within 2 seconds before and 4 seconds after the onset of the upstroke of the pharyngeal pressure wave.8 Multiple swallows were classified as a sequence of 2 or more swallows, which started within 2 seconds of the onset of the relaxation with an interswallow interval of less than 5 seconds. LESRs that occurred independently of swallowing (>2 seconds before or >4 seconds after pharyngeal contraction) with relaxation rates of ≥1 mm Hg/sec were classified as transient LES relaxations.8 Each LESR was also analyzed for resting pressure, nadir pressure, and duration of relaxation according to established criteria.2 Straining was identified by sustained increases (≥5 mm Hg) in intragastric pressure for at least 10 seconds, associated with a corresponding rise in esophageal body pressure. The computerized data analysis system made it possible to recognize the occurrence of 651

OMARI ET AL

THE JOURNAL OF PEDIATRICS NOVEMBER 1998

RESULTS Mechanisms and Rate of GER

Figure. Examples of manometric tracings of mechanisms of GER in premature infants. Spontaneous TLESR (A),TLESR occurring after esophageal body contraction (B), multiple swallows (C), and peristaltic failure (D). Dotted lines indicate onset of GER episode. Black bars indicate duration of LESR.

LESR during straining by comparison of the pressure differential between the gastric sensor and the sleeve.

IDENTIFICATION

OF

GER EPISODES.

Episodes of GER and associated manometric events were identified and classified. Esophageal pH changes and common cavity episodes (brief periods of gastroesophageal pressure equalization) were used as indicators of acid or nonacid (either milk or gas) reflux episodes. Acid GER was defined as an esophageal pH drop to ≤4 or by ≥1 pH unit for at least 5 seconds. Non-acid GER was detected by screening of the tracings for esophageal common cavity episodes2 that occurred without a significant pH drop. 652

ESOPHAGEAL CLEARANCE MECHANISMS. Acid GER episodes, identified by a drop in esophageal pH to below 4, were examined further to determine the effect of pressure wave patterning on the resolution of episodes of acidification. Acid clearance time was defined as the time taken for esophageal pH to return to a level of 4 or more.

Statistical Analysis Grouped data were compared by using an analysis of variance technique (F test and Scheffe’s test). Paired data were compared with a paired t test, and correlation was determined by simple regression analysis. A P value of < .05 was considered statistically significant.

Sixty acid and 133 non-acid GER episodes were recorded. The mean nadir pH of acid GER episodes was 3.0 ± 0.3 (range, 1.1 to 5.7) with a mean pH fall of 1.7 ± 0.1 (range, 1.0 to 2.9). Five infants had no acid reflux episodes. Only 30 of the 60 acid GER episodes recorded were accompanied by manometric recordings of sufficient resolution to identify common cavities. In this subgroup of events, common cavities were observed on 27 (90%) occasions. For all other acid GER episodes, the possible occurrence of a common cavity was obscured by pressures caused by movement artefact and/or straining. The data for the associations of reflux episodes with LES motor events and other variables are shown in Table I. Ninety-five percent of acid reflux episodes and all non-acid reflux episodes were associated with LESR. These were either TLESRs (Figure, A and B) or LESRs triggered by multiple swallowing (Figure, C), peristaltic failure (Figure, D), or vomiting (Table I). TLESRs triggered 74% of acid reflux episodes and 87% of non-acid reflux episodes (Table I). Thirty-one acid reflux episodes occurred during periods of abdominal straining, and TLESR was also the primary trigger for reflux. Only 2 reflux episodes occurred when LESP was detectable (>5 mm Hg), and both of these were in association with straining. The mean rate of combined acid/nonacid GER was 3.3 ± 0.4 episodes/h. Acid and non-acid GER occurred at frequencies of 1.0 ± 0.2 episodes/h and 2.3 ± 0.3 episodes/h, respectively. Because of the neutralizing effect of feeds (pH 7.5) on gastric acidity, acid GER rate was lower in the first postprandial hour compared with the second to third hours (0.6 ± 0.2 vs 1.2 ± 0.3 episodes/h, respectively, P < .05), whereas the non-acid GER rate was greatest in the first postprandial hour (2.9 ± 0.4 vs 1.8 ± 0.4 episodes/h, respectively, P < .05). There was no difference in GER rates between breast-fed and formula-fed infants (2.4 ± 0.5 vs 2.2 ± 0.4 episodes/h for non-acid GER and 0.7 ± 0.2 vs 1.4 ± 0.4 episodes/h for acid GER). Infants re-

THE JOURNAL OF PEDIATRICS VOLUME 133, NUMBER 5 ceiving theophylline or caffeine treatment had higher acid GER rates than untreated infants (1.7 ± 0.5 vs 0.7 ± 0.2 episodes/h, P < .05), but non-acid GER rates were similar (2.6 ± 0.7 vs 2.2 ± 0.4 episodes/h). The rate of acid GER did not correlate with PMA (r = 0.09, not significant); however, weak correlations were observed between lower PMA and a greater rate of both non-acid GER (r = 0.45, P < .05) and combined acid/nonacid GER (r = 0.46, P < .05), with younger infants demonstrating higher GER frequencies. The combined acid/non-acid GER rate correlated with the rate of TLESRs (r = 0.75, P < .0001).

Resting LES Pressure and Characteristics of LES Relaxation Mean resting LESP for individual infants ranged from 5.8 to 24.3 mm Hg (mean, 14.5 ± 1.1 mm Hg). There was no correlation between resting LESP and PMA (r = 0.32, not significant). Resting LESP did not differ between infants receiving expressed breast milk and those receiving formula during the study (14.9 ± 1.6 mm Hg vs 13.7 ± 1.7 mm Hg). Resting LESP in infants receiving theophylline or caffeine therapy did not differ from those not receiving this treatment (16.3 ± 1.5 mm Hg vs 13.7 ± 1.4 mm Hg). A total of 998 swallow-induced LESRs and 156 TLESRs were analyzed. TLESRs and multiple swallow-induced relaxations were longer in duration and achieved significantly lower nadir pressures when compared with single swallow-induced LESRs (durations were 15.1 ± 0.7 seconds and 10.8 ± 0.8 seconds vs 5.6 ± 0.4 seconds, respectively, P < .001; nadir pressures were 0.8 ± 0.2 mm Hg and 1.2 ± 0.2 mm Hg vs 2.3 ± 0.3 mm Hg, respectively, P < .001). Overall, TLESRs occurred at an average of 2.9 ± 0.3 per hour (range, 0.5 to 6.0 per hour). The mean rate of TLESRs was 3.5 ± 0.5 per hour in the first study hour and 2.3 ± 0.4 per hour during the remainder of the study (P = .07). There was a significant relationship between a higher TLESR rate and lower PMA (r = 0.61, P < .005). Infants receiving theophylline or caffeine treatment had overall TLESR rates similar to those of untreated infants (3.5 ± 0.6 vs 2.7 ± 0.4 episodes/h, P = .27);

OMARI ET AL

Table II. Esophageal pressure wave sequences occurring during periods of esophageal acidification

Esophageal pressure wave sequence

Swallowinduced (%)

Swallowunrelated (%)

Total (%)

Mean ∆ in pH

Peristaltic Synchronous Incomplete Retrograde None Total

143 (81.2) 13 (7.4) 14 (8.0) — 6 (3.4) 176 (100)

17 (8.0) 100 (47.4) 86 (40.8) 8 (3.8) — 211 (100)

160 (41.3) 113 (29.2) 100 (25.8) 8 (2.1) 6 (1.6) 387 (100)

0.59 ± 0.04 (0.0-2.9) 0.13 ± 0.03 (0.0-1.7)* 0.05 ± 0.02 (0.0-0.8)* 0.04 ± 0.03 (0.0-0.2)† 0.00 ± 0.00 (0.0-0.0)† 0.29 ± 0.02

Sequences occurred after 22 GER episodes during which esophageal pH dropped from above to below 4. The average increase in esophageal pH produced by this patterning is shown in the last column. Data are expressed as means ± SEM. *P < .0001. †P < .005 (∆ in pH significantly less than ∆ produced by peristaltic pressure wave sequences according to analysis of variance).

however, the treated infants had significantly higher rates of TLESRs in the first hour (5.3 ± 0.9 vs 2.7 ± 0.6 episodes/h, respectively, P < .05). Other characteristics of TLESR were unaltered by theophylline or caffeine treatment.

Esophageal Motor Patterns and Mechanisms of GER Clearance A total of 982 swallow-induced and 2234 swallow-unrelated esophageal body pressure wave sequences were analyzed. Seventy percent of all swallows triggered pressure waves that were peristaltic in sequence; of the remainder, 15% were synchronous, 11% were incomplete, and 4% failed to trigger a pressure wave sequence. In contrast, swallow-unrelated pressure waves were usually nonperistaltic (38% synchronous, 48% incomplete, and 9% retrograde) and rarely peristaltic (5%) in sequence. No correlations were found between the rate of pressure wave sequences and PMA. Twenty-two acid reflux episodes were identified by a pH drop from above to below 4. For these episodes, the mean clearance time was 193 ± 35 seconds (range, 16 to 510 seconds). The esophageal pressure wave sequences recorded during these periods of esophageal acidification were usually swallow-induced peristaltic or swallow-unrelated nonperistaltic (predominantly synchronous and incomplete) (Table II). Peristaltic pressure wave sequences (both swallow-induced and swallow-unrelated) produced significantly larg-

er increases in esophageal pH and therefore facilitated faster esophageal acid clearance than nonperistaltic pressure wave sequences (Table II).

DISCUSSION This study shows that in healthy preterm infants, TLESRs are the most common pattern of LES function associated with postprandial reflux. In these infants esophageal body motor patterns were well developed, and esophageal peristalsis effectively facilitated esophageal clearance of refluxate. We have previously shown that in premature infants TLESRs are the predominant mechanism of GER identified manometrically by common cavity phenomena (abrupt sustained pressurization of the esophageal body).2 In this study we have combined pH monitoring, the “gold standard” for recognition of acid GER, with manometry and have identified TLESRs as the single most important mechanism that allows acidic and non-acidic (liquid or gas) reflux of gastric contents across the LES in premature infants. This is now known to be the case in all age groups. In addition, LESR occurred appropriately with swallowing; and values for resting LESP, nadir LESP, and duration of LESR were consistent with our previously reported values.2,3 These data indicate that LES function is well developed, showing the same complex patterns 653

OMARI ET AL

of integrative function seen in older children and adults. TLESRs occur independently of swallowing and have a distinctive mechanical profile that is most likely determined by a hindbrain pattern generator. The current understanding of TLESRs and their control has been reviewed recently14; however, the neural mechanisms responsible for the initiation of TLESRs are still unclear. Gastric distension is thought to be a potent stimulus for initiation of TLESR,14 and this may explain the higher rate of TLESRs observed in the early postprandial period. Our data also indicate that TLESRs and total GER episodes occurred more frequently in younger infants. Conceivably, impaired gastric function in more premature infants could lead to delayed gastric emptying, prolonged gastric distension, and an increase in the frequency of TLESRs. Ewer et al15 have examined the relationship between gastric emptying rate and levels of GER (estimated by 24- hour pH monitoring) but were unable to determine a correlation. Clearly, the interrelationships among gastric emptying, the frequency of TLESR, and GER warrant further investigation. In our study feed type did not alter the rate of GER episodes; however, because most infants had received a variable mix of both formula and expressed breast milk in the postnatal period, it is difficult to provide an adequate commentary on the effect of feed type, which in other studies has been demonstrated to affect physiologic levels of GER.16 Infants receiving theophylline or caffeine treatment demonstrated higher frequencies of TLESRs and acid GER. The effect on acid GER is perhaps not surprising because xanthines have previously been shown to stimulate gastric acid secretion17 and in adults have been shown to augment GER and GER-related symptoms such as heartburn.18 Our data suggest that theophylline may also augment the frequency of TLESRs in the immediate postprandial period. However, because our study was not specifically designed to examine the impact of xanthine treatment and the numbers of infants in this group were relatively small (7 vs 16 control subjects), this observation will require further investigation. 654

THE JOURNAL OF PEDIATRICS NOVEMBER 1998 In our previous studies of esophageal body motor patterns in preterm infants,3 we found a high proportion of nonperistaltic pressure wave sequences, leading us to conclude that esophageal peristaltic mechanisms were poorly developed. However, our technique for recognition of swallowing was not adequate, and therefore we could not distinguish between swallow-induced and swallow-unrelated pressure wave sequences. Our current evaluation of motor responses to swallowing shows that 95% of swallows trigger esophageal body contractions that are predominantly (70%) peristaltic in sequence, indicating well-developed esophageal body motility. Esophageal pH monitoring showed that acid reflux episodes were predominantly cleared by swallow-induced, peristaltic esophageal body contraction sequences. Esophageal contractions that were nonperistaltic in sequence did not clear acid refluxate as effectively. If the relative proportion of nonperistaltic swallow-induced events were increased, this could lead to prolonged esophageal exposure to refluxate, which may be secondarily associated with obstructive apnea or aspiration and, in the longer term, with acid-induced mucosal damage. This study has shown that in healthy premature infants, as in adults and older children, the major mechanism underlying GER is TLESR. In addition, esophageal and LES motor function was well developed, even in premature infants of 31 weeks’ PMA. It is not known whether the same patterns of reflux and degree of maturation of esophageal body and LES function are present in premature infants younger than 31 weeks’ PMA and infants with clinically significant GER. Further studies utilizing the techniques described in this study are required to evaluate GER and esophageal clearance mechanisms in these infants. We thank Dr Kazu Miki for his participation in this study and Dr Charles Malbert for the computerized data acquisition and analysis software used in this study.

REFERENCES 1. Orenstein S. Gastroesophageal reflux. Curr Prob Pediatr 1991;21:193-241. 2. Omari T, Miki K, Davidson G, Fraser R,

3.

4.

5. 6.

7.

8.

9.

10.

11.

12.

13. 14.

15.

16.

17. 18.

Haslam R, Goldsworthy W, et al. Characterisation of lower oesophageal sphincter relaxation in healthy preterm infants. Gut 1997;40:370-5. Omari T, Miki K, Fraser R, Davidson G, Haslam R, Goldsworthy W, et al. Esophageal body and lower esophageal sphincter function in healthy premature infants. Gastroenterology 1995;109:1757-64. Newell S, Sarkar P, Durbin G, Booth I, McNeish A. Maturation of the lower oesophageal sphincter in the preterm baby. Gut 1988;29:167-72. Werlin S, Dodds W, Hogan W, Arndorfer R. Mechanisms of gastroesophageal reflux in children. J Pediatr 1980;97:244-9. Dent J, Holloway R, Toouli J, Dodds W. Mechanisms of lower oesophageal sphincter incompetence in patients with symptomatic gastroesophageal reflux. Gut 1988;29:1020-8. Dent J, Dodds W, Friedman R, Sekiguchi T, Hogan WJ, Arndorfer RC, et al. Mechanism of gastroesophageal reflux in recumbent asymptomatic human subjects. J Clin Invest 1980;65:256-67. Holloway R, Penagini R, Ireland A. Criteria for the objective definition of transient lower esophageal sphincter relaxation. Am J Physiol 1995;268:G1-8. Cucchiara S, Bortolotti M, Minella R, Auricchio S. Fasting and postprandial mechanisms of gastroesophageal reflux in children with gastroesophageal reflux. Dig Dis Sci 1993;38:86-92. Schoeman M, Holloway R. Integrity and characteristics of secondary oesophageal peristalsis in patients with gastro-oesophageal reflux disease. Gut 1995;36:499-504. Kawahara H, Dent J, Davidson G. Evaluation of the mechanisms responsible for gastroesophageal reflux in children. Gastroenterology 1997;113:399-408. Omari T, Bakewell M, Fraser R, Davidson G, Dent J. Intraluminal micromanometry: an evaluation of the dynamic performance of micro-extrusions and sleeve sensors. Neurogastroenterol Motil 1996;8:241-5. Dent J. A new technique for continuous sphincter pressure measurement. Gastroenterology 1976;71:263-7. Mittal R, Holloway R, Penagini R, Blackshaw L, Dent J. Transient lower esophageal relaxation. Gastroenterology 1995;109:601-10. Ewer A, Durbin G, Morgan M, Booth I. Gastric emptying and gastro-oesophageal reflux in preterm infants. Arch Dis Child 1996;75:F117-21. Heacock H, Jeffery H, Baker J, Page M. Influence of breast versus formula milk on physiological gastroesophageal reflux in healthy, newborn infants. J Pediatr Gastroenterol Nutr 1992;14:41-6. Forster LJ, Trudeau WL, Goldman AI. Bronchodilator effects on gastric acid secretion. JAMA 1979;241:2613-5. Berquist WE, Rachelefsky GS, Kadden M, Siegel SC, Katz R, Fonkalsrud EW, et al. Effect of theophylline on gastroesophageal reflux in normal adults. J Allergy Clin Immunol 1981;67:407-11.