Objective. To examine the temporal relationship between apnea and gastroesophageal reflux (GER) and to assess the effect of GER on apnea duration.
Methods. A total of 119 preterm infants underwent 12-hour cardiorespiratory monitoring studies using respiratory inductance plethysmography, heart rate, oxygen saturation (SaO2), and esophageal pH. The studies were scored for GER (pH <4 for ≥5 seconds) and apnea ≥15 seconds or ≥10 seconds that occurred within 30 seconds of GER. Apnea ≥10 seconds was used to assess whether GER would prolong apnea duration.
Results. There were 6255 episodes of GER. Only 1% of GER episodes were associated with apnea ≥15 seconds, and there was no difference in apnea rate before, during, or after GER. There was also no difference in rate of apnea ≥10 seconds before versus during GER; however, there was a decrease in apnea rate immediately after GER. The presence of GER during apnea did not prolong apnea duration, and GER had no effect on the lowest SaO2 or heart rate during apnea.
Conclusion. There is no evidence of a temporal relationship between acid-based GER and apnea in preterm infants. In addition, GER does not prolong apnea duration and does not exacerbate the resultant decrease in heart rate and SaO2.
Gastroesophageal reflux (GER) and apnea are 2 clinical events that both occur commonly in the premature infant population. Early studies have suggested a relationship between GER and apnea1–4 in term and preterm infants, possibly contributing to widespread use of antireflux medications to treat apnea of prematurity. More recent studies have consistently failed to support this relationship5–9; however, the use of antireflux medications in the treatment for apnea of prematurity remains widespread.
Although a significant overall temporal relationship between apnea and GER has not been established, these same studies have shown that up to 19% of apneic episodes may occur concurrently with GER.5,7,10 It is unclear, in such instances, whether short respiratory pauses may be prolonged when accompanied by GER, thereby becoming clinically significant events.
Therefore, the purpose of this study was to examine the temporal relationship between apnea and GER in a large group of premature infants to assess the effect of GER on apnea duration. We hypothesized that when GER and apnea do coincide, GER would prolong the duration of apnea. If short respiratory pauses are indeed significantly prolonged when accompanied by GER, then this might justify the use of antireflux medications in the treatment of such episodes of apnea.
We reviewed 12-hour overnight bedside cardiorespiratory monitoring studies in a population of 119 premature infants at Rainbow Babies & Children's Hospital recorded over a period of 5 years (1998–2002) with demographic data presented in Table 1. Infants who were enrolled in the study (97% of whom were <1500 g at birth) were of comparable weight and gestational age as infants who were <1500 g at birth and were not enrolled in the study during the same period (1019 ± 301 g vs 1060 ± 273 g and 28 ± 2 weeks vs 28 ± 3 weeks, infants in the study versus infants not enrolled in the study, respectively). All infants were referred for an overnight bedside cardiorespiratory monitoring study by their attending physician because of suspected clinical symptoms of GER, with or without persistence of apnea, bradycardia, and/or desaturation. Exclusion criteria included the presence of major congenital anomalies. Administration of supplemental oxygen, antireflux medications, and xanthine therapy were noted (Table 1). Standard doses of theophylline and caffeine were used, and theophylline blood levels were not typically available at the time of study. As some infants who were receiving antireflux medications still had acid-based GER, these infants were not excluded from the study. All infants were receiving full feeds at intervals of 2 hours (7% of infants), 3 hours (61% of infants), or 4 hours (32% of infants). Because the infants were monitored overnight at the bedside, body position varied during the study per routine clinical practice. Institutional Review Board approval was obtained, and informed consent was acquired.
Each 12-hour overnight cardiorespiratory study consisted of respiration measured via respiratory inductance plethysmography, oxygen saturation (SaO2), pulse wave form, heart rate (SomnoStar PT; Sensormedics, Yorba Linda, CA), and esophageal pH. Respiratory inductance plethysmography entailed 2 coil bands placed around the rib cage and abdomen. Rib cage and abdominal excursions then were summed to acquire an estimate of tidal volume using the Qualitative Diagnostic Calibration procedure previously described by Sackner et al.11 This procedure yields a semiquantitative value for tidal volume expressed as units of percentage tidal volume during the monitor calibration period. A valid breath was defined as a tidal volume ≥25% of the calibration period. A semidisposable monocrystant antimony pH catheter attached to a pH meter (Synectics, Stockholm, Sweden) was used to record esophageal pH. The pH probe was placed in the lower third of the esophagus between T7 and T9, with its position confirmed radiographically.
Apneic events were scored during periods when respiratory patterns could be distinguished from motion artifact, defined as gross erratic movements on multiple wave forms including the respiratory channels. Apnea was defined as cessation of breathing for ≥15 seconds and ≥10 seconds. The latter was used to address the effect of GER on duration of shorter respiratory pauses. The duration of the respiratory pause was recorded from the tidal volume channel. Central and obstructive components for each event were also recorded along with the lowest heart rate and SaO2. The obstructive component was defined as the occurrence of asynchronous rib cage and abdominal movement during periods when a valid breath (≥25% of the baseline period) was not present on the tidal volume wave form. Apnea was subsequently categorized as central, obstructive, or mixed, with mixed apnea containing a central pause and at least 1 obstructive effort. Episodes of hypopnea were not quantified for this study.
SaO2 values that were contaminated by motion artifact were removed from the data analysis. Motion artifact was identified by erratic baseline shifts on the pulse wave form where a clear pulse for every R wave on the electrocardiogram could not be distinguished from the pulse wave form. Episodes of desaturation (SaO2 ≤85%) and bradycardia (≤80 beats per minute) were also noted. Studies were scored visually using RespiEvents Software (Sensormedics). Because this was an unattended study, it was not possible to distinguish sleep state in this population.
The 12-hour studies were screened for any episodes of acid GER (pH <4 for ≥5 seconds), and a reflux index, defined as the percentage of the study with pH <4, was recorded for each infant. Once an episode of GER was identified, any apnea with its onset occurring within 30 seconds before GER, the entire period during GER, or within 30 seconds after GER was recorded. Apnea that extended between these time windows was classified in the window in which the apnea began. Because the events were identified by the occurrence of GER and apnea was scored only during and within 30 seconds of GER, it was not possible to blind the scorer as to the presence of GER.
During reflux episodes that were associated with apnea, apnea rate was compared between the 30-second period before, the entire period during, and the 30 seconds after resolution of GER. Because a number of the GER episodes occurred over prolonged periods when the frequency of apnea may have changed, apnea frequency was also calculated for the first 30 seconds of the reflux episode. For accounting for varying GER duration, all apnea rates were normalized as number of apnea/30 seconds. For determination of whether GER may have initiated a behavioral arousal or startle, thus potentially affecting apnea rate, any body movement that occurred within 30 seconds of the termination of the reflux episode was noted. A behavioral arousal or startle was defined as any gross erratic movement on the electrocardiogram, pulse, and/or respiratory wave form.
To assess whether initiation of GER would prolong apnea duration, we examined infants who had both apnea of ≥10 seconds that occurred during the baseline period that preceded GER with resolution before the initiation of GER and apnea that occurred during the baseline period and continued through initiation of GER. One-way analysis of variance (ANOVA) with repeated measures and post hoc analysis via Student-Newman-Keuls' multiple-range test were used to compare apnea rate before, during, and immediately after an episode of GER. Paired t test was used to assess the effect of GER on apnea duration. A post hoc power analysis for apnea duration was performed on the basis of the data available in 27 of the 119 infants who were enrolled in this study, comprising infants in whom apnea occurred during the baseline period that terminated both before initiation of GER and after initiation of GER. This revealed the study to have a power of 80% to detect a 25% increase in apnea duration in response to GER with α = .05. P < .05 was needed for significance.
There were 6255 episodes of GER, or 53 ± 41 episodes/12-hour study. The mean duration of GER was 2.25 ± 1.97 hours/12-hour study with a reflux index of 17 ± 14% (median: 13%; range: 0%–69%). Treatment with xanthines (n = 26) during the study had no effect on the reflux index (17 ± 17% vs 17 ± 14%, xanthine vs no xanthine therapy, respectively; P = .94)
Temporal Relationship of Apnea ≥15 Seconds to GER
Twenty-seven (23%) of 119 infants had at least 1 episode of reflux associated with apnea ≥15 seconds, with a total of 64 of 6255, or 1%, of acid-based reflux episodes being associated with apnea ≥15 seconds. Of these 64 episodes of apnea associated with GER, 74% were central, 21% were mixed, and 5% were obstructive.
Comparison of apnea rates associated with acid-based reflux episodes for these 27 infants is shown in Fig 1. There was no difference in frequency of apnea before, during, or after GER episodes (P = .30, ANOVA), although there was a trend toward a decrease in apnea frequency during resolution of GER. Fifty-two (81%) of the 64 reflux episodes that were associated with apnea had a behavioral arousal within 30 seconds of the termination of the reflux episode, which may have contributed to the decrease in apnea rate.
Temporal Relationship of Apnea ≥10 Seconds to GER
Eighty-three of 119 infants had at least 1 episode of reflux associated with apnea ≥10 seconds, with a total of 532 of 6255 (8.5%) of all reflux episodes being associated with apnea ≥10 seconds. Of these 532 episodes of apnea associated with GER, 86% were central, 12% were mixed, and 2% were obstructive.
Comparison of apnea rates associated with reflux episodes for the 83 infants is shown in Fig 2. Although there was no difference in apnea rate before versus during GER, there was a significant decrease in frequency of apnea immediately after GER episodes when compared with apnea frequency both before and during GER (P = .01, ANOVA, and P = .05 after vs before and during GER, respectively, Newman Keuls'). There was a higher incidence of mixed/obstructive apnea during GER with 5 ± 15% versus 11 ± 26% of apneic episodes being mixed or obstructive before versus during GER, respectively (P = .017, t test). A total of 450 of the 532 (85%) reflux episodes that were associated with apnea had a behavioral arousal within 30 seconds of the termination of the reflux episode.
Effect of GER on Apnea Duration
We sought to assess whether initiation of GER during apnea would prolong apnea duration. Therefore, we examined 27 of the 83 infants who had both apnea that occurred during the baseline period that preceded GER with resolution before the initiation of GER and apnea that occurred during the baseline period that continued through initiation of GER. GER that began during an apnea had no significant effect on apnea duration, with the duration of apnea that resolved before initiation of GER being 12.3 ± 2.2 seconds in contrast to apnea that continued through initiation of reflux having a duration of 13.9 ± 6.9 seconds (P = .17, paired t test; Fig 3). In addition, GER had no effect on either the lowest SaO2 or heart rate during apnea (SaO2: 87 ± 6 vs 87 ± 11%, P = .9; heart rate: 125 ± 25 vs 129 ± 26 per minute, P = .4 for apnea that resolved before versus during GER, respectively).
The data presented in this study indicate no temporal relationship between acid GER and apnea. This lack of temporal relationship is in agreement with previous studies5–7 that comprised smaller infant populations that ranged from 19 to 45 infants. The current study is based on data derived from 119 infants, >6000 episodes of GER, and a subset of 83 infants with GER associated with apnea. This has enabled us to demonstrate for the first time that in the few instances when GER occurs concurrently with apnea, GER does not prolong apnea duration.
Laryngeal stimulation has been shown to induce apnea in piglets.12–14 We therefore hypothesized that initiation of GER during apnea might stimulate superior laryngeal nerve or other afferents and thus prolong apnea duration. In contrast to our hypothesis, GER had no effect on apnea duration. It may be that the acid bolus detected by pH monitoring in the esophagus was present only in the lower esophagus (mild reflux) and did not reach the larynx or pharynx (severe reflux). It may also be possible that it did reach the larynx but did not elicit the threshold for laryngeal inhibition.
Seventy-four percent of apneas were central in origin. Although the majority of apneas were central, there was a higher percentage of mixed/obstructive apnea during GER when compared with the period before GER, which is in agreement with previous findings by Walsh et al.15 Menon et al4 noted frequent swallowing during prolonged apnea after regurgitation in infants. It is possible that GER may precipitate swallowing as a clearing mechanism with an accompanying brief period of airway closure. This may predispose the infant to obstructive breaths as respiratory efforts resume.
Previous studies in preterm infants have reported inconsistent results pertaining to the incidence of central versus mixed/obstructive apnea.7,15–21 Although RIP technology may potentially miss intermittent obstructive breaths,22 the majority of these studies have found similar incidences of central apnea with ranges of 50% to 69%15,16,18–20 using multiple techniques for apnea detection, including RIP, impedance, thermistor, pneumotachograph, and pressure transducer. In addition, Brooks et al23 found good agreement between RIP and volume measurements made via nasal mask pneumotachograph, the gold standard for volume measurements in infants. Although underestimation of obstructive efforts may have affected our ability to quantify apnea duration, our technology nonetheless was able to detect a higher incidence of obstructive breaths during episodes of GER. Future studies in a sleep laboratory environment using more invasive technology would be needed to address this issue.
There was a decrease in the occurrence of apnea at resolution of GER. This corresponded with the high incidence of a behavioral arousal or startle, associated with termination of the reflux episode. Behavioral arousal was shown previously to be associated with, although not essential for, resolution of apnea.16 Alternatively, apnea has been shown to be initiated by an arousal.24 Because more subtle measurement of arousal using electroencephalogram and video-based documentation of behavioral changes was not available for this study, the potential effect of arousal or startle on apnea rate is unclear.
In this study, GER was quantified using pH monitoring. The current methods for measuring reflux entail pH monitoring and the intraluminal impedance technique. As pH monitoring will detect only acid-based reflux (pH <4), this technique may underestimate the occurrence of reflux when compared with intraluminal impedance, which is pH independent and therefore will detect any bolus present in the esophagus.25 However, even with the potential underestimation of GER, there are no current data to suggest that non–acid-based reflux would behave differently. Furthermore, the findings of this study are consistent with Peter et al,5 who found no temporal relationship between apnea and both acid- and non–acid-based GER using the multiple intraluminal impedance technique in preterm infants.
It is unclear whether potential discrepancies in measurement of reflux may be compounded further by feeding intervals.26,27 Hegar et al27 found no difference in the reflux index when using the entire study versus exclusion of postprandial periods or all periods with gastric pH >4.0. After feeds, stomach acid is neutralized, thereby possibly limiting detection of reflux by pH monitoring, although recent data indicate that previous studies overestimated the effect of feeding on gastric acidity.28 Body position is another variable that may affect GER,29 which was not controlled for in this study. Xanthines have been shown to decrease occurrence of apnea and increase the incidence of GER30; however, xanthine therapy had no effect on the reflux index in these infants. This may be attributable to the small number of infants (n = 26) who were receiving xanthine therapy at the time of study. Even with the potential confounding factors of pH monitoring, xanthine therapy, feeding intervals, and body position, the median reflux index of 13% is comparable to previous studies in symptomatic preterm infants,7,29 with reflux indices ranging from 13.8% to 16%. In addition, although these confounding factors may have affected the incidence of GER and/or apnea, they should not have affected the relationship between the 2 events.
In conclusion, we have shown no evidence for an overall temporal relationship between acid-based GER and apnea in preterm infants. We have demonstrated that, in the few instances in which GER occurs concurrently with apnea, GER has no effect on apnea duration. It is possible that a small subgroup of infants may respond favorably to antireflux medications. However, these data suggest that, in the general premature infant population, GER does not induce apnea, does not prolong apnea duration, and does not exacerbate the apnea-related decrease in heart rate and SaO2. Therefore, we cannot support the continued widespread use of antireflux medications for the treatment of apnea of prematurity.
- Accepted January 24, 2005.
- Reprint requests to (J.M.D.F.) Department of Pediatrics, Division of Neonatology, Rainbow Babies & Children's Hospital, 11100 Euclid Ave, Cleveland, OH 44106. E-mail:
No conflict of interest declared.
- Leape LL, Holder TM, Franklin JD, Amoury, RA, Ashcraft KW. Respiratory arrest in infants secondary to gastroesophageal reflux. Pediatrics.1977;60 :924– 928
- ↵Peter CS, Sprodowski, N, Bohnhorst, B, Silny J, Poets C. Gastroesophageal reflux and apnea of prematurity: no temporal relationship. Pediatrics.2002;109 :8– 11
- ↵Sackner MA, Watson H, Belsito AS, et al. Calibration of respiratory inductive plethysmograph during natural breathing. J Appl Physiol.1989;66 :410– 420
- ↵Lawson EE. Prolonged central respiratory inhibition following reflex-induced apnea. J Appl Physiol. 198;50 :874– 879
- Haxhiu-Poskurica B, Carlo WA, Miller MJ, Di Fiore JM, Haxhiu MA, Martin RJ. Maturation of respiratory reflex responses in the piglet. J Appl Physiol.1991;70 :608– 616
- ↵Thoppil CK, Belan MA, Cowen CP, Mathew OP. Behavioral arousal in newborn infants and its association with termination of apnea. J Appl Physiol.1991;70 :2479– 2484
- ↵Di Fiore JM, Arko MK, Miller MJ, et al. Cardiorespiratory events in preterm infants referred for apnea monitoring studies. Pediatrics.2001;108 :1304– 1308
- ↵Mathew OP, Thoppil CK, Belan M. Motor activity and apnea in preterm infants. Am Rev Respir Dis.1999;144 :842– 844
- ↵Grant L, Cochran D. Can pH monitoring reliably detect gastro-oesophageal reflux in preterm infants? Arch Dis Child Fetal Neonatal Ed.2001;85 :F155– 157
- ↵Omari TI, Davidson GP. Multipoint measurement of intragastric pH in healthy preterm infants. Arch Dis Child Fetal Neonatal Ed.2003;88 :F517– F520
- ↵Ewer AK, James ME, Tobin JM. Prone and left lateral positioning reduce gastro-oesophageal reflux in preterm infants. Arch Dis Child Fetal Neonatal Ed.1999;81 :F201– F205
- ↵Newell SJ, Booth IW, Morgan ME, Durbin GM, McNeish AS. Gastro-oesophageal reflux in preterm infants. Arch Dis Child.1989;64 :780– 786
- Copyright © 2005 by the American Academy of Pediatrics