Snel, Antonie; Barnett, Christopher P.; Cresp, Trevor L.; Haslam, Ross R.; Davidson, Geoffery P.*; Malbert, Charles H.†; Dent, John‡; Omari, Taher I.*
Neonatal Medicine and *Gastroenterology Unit, Women's and Children's Hospital, North Adelaide, Australia; †Unite Des Flux Digestif, Institute National de la Recherche Agronomique, St. Gilles, France; and ‡Gastrointestinal Medicine, Royal Adelaide Hospital, Adelaide, Australia
Received March 25, 1999;
revised June 14, 1999; accepted July 19, 1999.
Supported by the National Health and Medical Research Council of Australia, the J. H. and J. D. Gunn Medical Research Foundation, and Channel 7 Children's Research Foundation of South Australia.
Address correspondence and reprint requests to Dr. Taher Omari, Gastroenterology Unit, Women's and Children's Hospital, North Adelaide, SA 5006, Australia.
This article is accompanied by an editorial. Please see: Rudolph CD. Probing questions: When is gastroesophageal reflux the cause of symptoms? J Pediatr Gastroenterol Nutr 2000;30;3–4.
Gastroesophageal reflux (GER) occurs commonly in premature babies. Frequent regurgitation or vomiting after feeding is the most easily recognized clinical sign of pathologic reflux (GER disease) which may result in weight loss or failure to thrive. Frequent GER increases the likelihood of aspiration, which can cause acute respiratory distress (aspiration pneumonia) or exacerbate chronic lung disease. Gastroesophageal reflux may also be secondarily associated with obstructive apnea and bradycardia (1,2).
In adults and older children, 24-hour pH monitoring has been an effective clinical tool for the recognition of potentially pathologic degrees of esophageal acid exposure (3). These patients are also usually able to report symptoms such as chest pain, nausea, or heartburn to their physicians, enabling the association between symptoms and the occurrence of acid reflux episodes to be made. In the same way, the effectiveness of therapy is readily assessed. In premature infants, basal acid secretion reaches levels equivalent to those of adults and children by 4 weeks of postnatal age (4). pH-monitoring studies in premature infants (5,6) have indicated that acid GER is a common event in these babies and, interestingly, the levels of esophageal acidification (indicated by reflux index scores) are higher on average than the levels considered clinically significant in adults and children. The inability of neonates to report their symptoms and the absence of normal pH-monitoring data in this age group make the accurate diagnosis of GER disease in premature infants more difficult. Indeed, there are currently no consistent gold standard clinical criteria for GER disease in these infants.
The ability to recognize reflux-specific behavioral patterns in premature infants would probably be the most sensitive means of clinical diagnosis of reflux disease, and empirically, most treatment decisions in premature infants are based on the presence of specific symptoms. Nonspecific changes in infant behavior that are believed to indicate discomfort, such as irritability, crying, or grimacing, are often observed by nursery staff, especially during feeding and in the early postprandial period. Such behavior may reflect pain or discomfort associated with esophageal acidification or the noxious acidic taste of refluxate. This is certainly the case in older infants with reflux disease (7). The purpose of this study was to use concurrent esophageal pH monitoring and video imaging to investigate whether major esophageal acidification caused by episodes of GER alters the behavioral patterns of healthy premature infants.
Studies were performed in 14 preterm infants (9 male, 5 female) born at a mean gestational age of 31 weeks (range, 26–35 weeks) with a mean postnatal age at the time of study of 21 days (range, 13–79 days). The mean corrected gestational age (gestational age + postnatal age) was 35 weeks (range, 32–38 weeks). All infants were well at the time of the study, were considered clinically free of reflux disease, had no evidence of neurologic dysfunction, and were not receiving prokinetic medication. Mean infant weight was 2150 g (range, 1720–2800 g). Seven infants were receiving either theophylline or caffeine treatment because of previous episodes of apnea and bradycardia. Eight infants were gavage fed with nonfortified expressed breast milk and six with infant formula (Enfalac 20/24 calorie; Mead Johnson, Canada).
The study protocol was approved by the Ethics Research Committee of the Women's and Children's Hospital, and written informed parental consent was obtained before each study.
pH and Manometric Recording Technique
Esophageal pH was monitored with an antimony pH probe (outside diameter, 1.5 mm; Synectics, Stockholm, Sweden) that was attached with rings of thin elastic silicone rubber to a manometric assembly (outside diameter, 2 mm) that was used both to administer feedings and to record pressures, so that the pH probe was accurately positioned within the esophageal body.
Analog pressure-transducer and pH-probe signals were amplified and filtered with a polygraph (Synectics) and digitized at 10 Hz with an analog-to-digital card (National Instruments, Austin, TX, U.S.A.). Data acquisition and analysis were performed on a computer (Macintosh Quadra 700; Apple Computer, Cupertino, CA, U.S.A.) with software based on National Instruments' Labview (MAD software, Royal Adelaide Hospital, C. Malbert). Infant behavior was continuously recorded at bedside with a color video camera (WV-GL350; Panasonic, Australia) coupled to a video cassette recorder (AG-6500 National Instruments). Video images and the pH and manometry recordings were synchronized by a video time coder (TD-100S, Provideo Systems Pty. Ltd., Norwood, South Australia).
The assembly was passed transnasally and positioned with the pH probe 3 cm above the manometrically identified lower esophageal sphincter high-pressure zone. After positioning the assembly, the feed was administered during a 15-to 30-minute period through the core channel (inside diameter, 0.75 mm) of the assembly. Esophageal pH and behavioral patterns were recorded during feeding and then for 3 to 4 hours after the meal.
Identification of Gastroesophageal Reflux Episodes
Episodes of acid GER were identified by rapid decreases in esophageal pH that lasted at least 15 seconds. For the purpose of this analysis, only episodes characterized by major acidification were examined. These were defined as an esophageal pH decrease from above 4 to below, in which total change in pH was more than 1 pH unit (i.e., the nadir pH during the episode was always <4).
Analysis of Video Recordings
For each episode of major acid GER identified, a 10-minute video epoch, encompassing the onset of acid GER and at least 4 minutes after GER, was transferred to a separate video tape for later analysis. In addition, to serve as a control, a second 10-minute epoch, occurring within ±30 minutes of the GER episode and during which basal esophageal pH was always more than 4, was transferred to a separate video tape. The 10-minute epochs were then randomized and viewed independently by two observers (T.C. and C.B.) who were blinded to the pH tracings.
Both observers scored the video recordings for general behavior and reflux-specific behavior according to criteria established by Davidson et al. (8) and Feranchak et al. (7) General behavior, consisting of resting with eyes open, resting with eyes closed, moving and comfortable, restless and uncomfortable, and irritable and crying, was scored continuously at 15-second intervals (i.e., 40 times during 10 minutes). Reflux-specific behavior, consisting of discomfort, regurgitation or spilling, yawning, mouthing, hiccupping, sneezing, coughing and gagging, thumb sucking, and head retraction, was scored when applicable. For the purposes of analysis, the behavioral patterns scored during the 4-minute period immediately after GER were directly compared with those scored during the equivalent 4 minutes of the control period.
Agreement of behavioral scoring between the two observers was assessed using Cohen's κ coefficient which was compared with the standard two-tailed normal distribution to determine significance of agreement. Behavioral data were compared nonparametrically by Wilcoxon's signed-rank test. P < 0.05 was considered statistically significant.
The combined manometric–pH-monitoring–feeding assembly was well tolerated by all infants, and no adverse effects related to the procedure were noted. Twenty episodes of acid GER were recorded and analyzed. One acid GER episode was characterized in each of 10 infants, two episodes were characterized in 2 infants, and three episodes were characterized in the remaining 2 infants. These occurred between 52 and 199 minutes after a meal, with a pH decrease ranging from 1.0 to 4.5 pH units (mean decrease, 2.2) and a nadir pH ranging from 0.8 to 3.8 (mean nadir pH, 2.5). No problems were encountered with the behavioral scoring system. There was a 74% agreement of behavior scoring by the two observers. The κ coefficient was 0.488 ± 0.014 (Z = 34.8;P < 0.001) indicating good interobserver agreement.
General Behavioral Patterns
The premature infants studied were usually observed resting with eyes open. Other behavioral patterns were less common, particularly those considered to be indicative of pain, such as restless and uncomfortable and irritable and crying behavior. The presence of acid reflux did not alter the pattern of general behavior when compared with control. A statistically significant reduction in resting with eyes closed was observed; however, this was scored by only one of the observers (Fig. 1).
Both observers scored discomfort, mouthing, and head retraction more frequently than other reflux-specific behavior. Regurgitation or spilling, yawning, hiccupping, sneezing, and thumb sucking were rarely observed, and coughing or gagging was never recorded. For both observers, the presence of acid reflux did not significantly alter the scoring of reflux-specific behavioral patterns (Fig. 2).
There are no previously reported data on the behavioral responses of premature infants to acid GER. Despite this, it is common nursery practice to use behavioral characteristics such as discomfort, head retraction, and mouthing in the diagnosis and subsequent management of GER in these infants. Using two blinded observers, we have been able to show that healthy premature infants frequently demonstrate behavior that, in symptomatic older infants, is associated with the occurrence of acid reflux, but that this behavior is not associated with reflux in healthy premature infants.
Given the paucity of published data on GER-related behavior in preterm infants, in our study we used the behavioral markers characterized by Feranchak et al. (7) in term infants under video surveillance during acid exposure. In that study, six discrete types of behavior were identified that were clearly associated with the onset of reflux events in infants with GER disease referred for 24-hour esophageal pH monitoring. The premature infants enrolled in our study frequently demonstrated reflux-specific behavior, including discomfort, head retraction, and mouthing; however, none of these behavioral patterns was temporally associated with the occurrence of acid GER. Our data also indicate that general behavioral criteria such as discomfort may simply be too difficult to detect readily in premature infants. Instead, responses to GER may have been reported more frequently if discomfort was scored using a general neonatal pain score such as the Neonatal Facial Coding System (11).
The premature infants in our study were, to the best of our knowledge, healthy and free of reflux disease. Nevertheless, it was surprising that esophageal acidification did not elicit any behavioral response at all. Indeed, on review of the video tapes at the conclusion of the study, infants were often motionless and without any facial expression, even after prolonged periods of esophageal acid exposure. Because premature infants have clear responses to painful stimuli (9,10) and the levels of physiological acid reflux are higher in premature infants than in older children and adults (5,6), our data suggest that relatively infrequent but prolonged episodes of esophageal acid exposure are well tolerated by healthy premature infants and so do not necessarily lead to a perceptible sensation from the esophagus.
Our study was designed to identify unequivocal esophageal acidification episodes (mean nadir pH, 2.5) in the hope of identifying concurrent behavioral patterns. Our data indicate that healthy premature infants often demonstrate behavioral patterns that, in older infants with GER disease, have previously been associated with the occurrence of GER. However, we were unable to identify a temporal association between the observation of reflux-specific behavior and the occurrence of esophageal acidification. Our findings clearly indicate a need for the establishment of reflux-related behavioral criteria that are specific to premature infants, because the criteria established by Feranchak et al. (7) are likely to be misleading when used in this age group. Application of these criteria to preterm neonates is likely to lead to an incorrect diagnosis of GER and unnecessary use of antireflux therapy.
1. Orenstein S, Orenstein D. Gastroesophageal reflux and respiratory disease in children. J Pediatr 1988; 112:847–58.
2. Orenstein S. Gastroesophageal reflux. Curr Probl Pediatr 1991; 21:193–241.
3. Vandenplas Y, Sacre–Smits L. Continuous 24-hour esophageal pH monitoring in 285 asymptomatic infants 0–15 months old. J Pediatr Gastroenterol Nutr 1987; 6:220–4.
4. Hyman P, Clarke D, Everett S, et al. Gastric acid function in preterm infants. J Pediatr 1985; 106:467–71.
5. Jeffery H, Page M. Developmental maturation of gastro-oesophageal reflux in preterm infants. Acta Paediatr. 1995; 84:245–50.
6. Newell S, Booth I, Morgan M, Durbin G, McNeish A. Gastro-oesophageal reflux in preterm infants. Arch Dis Child. 1989; 64:780–6.
7. Feranchak A, Orenstein S, Cohn J. behaviours associated with onset of gastroesophageal reflux episodes in infants. Clin Pediatr 1994;654–61.
8. Davidson G, Dent J, Willing J. Monitoring of upper oesophageal sphincter pressure in children. Gut 1991; 32:607–11.
9. Blauer T, Gerstmann D. A simultaneous comparison of three neonatal pain scales during common NICU procedures. Clin J Pain 1998; 14:39–47.
10. Shah V, Taddio A, Bennett S, Speidel B. Neonatal pain response to heel stick vs venepuncture for routine blood sampling. Arch Dis Child Fetal Neonatal Ed 1997; 77:F143–4.
11. Grunau R, Johnson C, Craig K. Neonatal facial and cry response to invasive and non-invasive procedures. Pain 1993; 52:201–8.
© 2000 Lippincott Williams & Wilkins, Inc.