Effect of Severity of Esophageal Acidification on Sleep vs. Wake Periods in Infants Presenting with Brief Resolved Unexplained Events

Abstract

Objectives

To describe the pattern of gastro-esophageal reflux (GER) events in wake and sleep states with increasing acid reflux index (ARI) in neonates and to test the hypothesis that GER-related symptoms are frequent in ARI> 7% in wake-state.

Study design

Infants underwent 24-hour pH-impedance studies with 6-hour concurrent video-polysomnography studies. Data were stratified based on the 24-hr ARI (% duration that esophageal pH is <4) into ARI<3% (normal), ARI 3≥ to ≤7% (intermediate) and ARI>7% (abnormal). GER frequency, clearance mechanisms and symptoms were distinguished during wake-state and sleep-state.

Results

Total wake and sleep duration was similar (p≥ 0.2) in all ARI groups. Acidic events were frequent with increasing ARI in wake-state vs. sleep-state (p≤ 0.03). The Symptom Index increased with increasing ARI (p≤ 0.02) in both wake-state and sleep-state. Acid clearance time increased with increasing ARI in wake-state (p≤ 0.02). In ARI>7% vs. ARI≤7%, frequency of acidic GER events was higher (p≤0.02) in wake-state and sleep-state; proximal migration of acid (p=0.03) and acid clearance time were higher in wake-state (p=0.0005) only. Symptom index was higher in ARI>7% vs. ARI≤7% in wake-state (p<0.0001), comparable in normal vs. intermediate (p=0.4) and higher in abnormal vs. intermediate (p=0.0004) groups.

Conclusions

Severe esophageal acid exposure (ARI>7%) is associated with increased reflux-associated symptoms in wake-state. Sleep-state appears to be protective regardless of ARI, likely due to greater chemosensory thresholds. Attention to posture and movements during wake-state can be helpful. Scrutiny for non-GER etiologies should occur for infants presenting with life-threatening symptoms.

More than 50% of infants younger than 3 months have some degree of gastro-esophageal reflux (GER) symptoms,¹ and the prevalence of GER disease in neonatal intensive care unit (NICU) infants is around 10%.² Magnitude of acidity (acid- and weakly-acid exposure) and severity of esophageal acid exposure as measured by acid reflux index (ARI) are important considerations in the genesis of symptoms.³ Acid reflux index is defined as the percentage of the entire pH recorded time that esophageal pH is <4.¹ ARI < 3% is considered normal, >7% is abnormal and ARI 3≥ to ≤7% is considered intermediate.¹ ARI is reportedly an effective tool in predicting prognosis; symptoms at 1-year follow up were higher in infants with high initial ARI.⁴

Sleep is a complex amalgam of modifiable physiological and behavioral processes.^5,6 Neurosensory processing and learning that are maximum during quiet sleep are disrupted by GER.⁷ Esophageal distention or chemo-sensitive stimulation during GER events result in esophageal, pharyngeal, glottal and laryngeal reflexes that protect the aerodigestive tract and prevent aspiration,^8–10, and are based in part on the degree of acidity of the reflux event.¹¹ However, the role of sleep or wake state on these mechanisms is not entirely clear.

Cardiorespiratory and physical symptoms have been associated with GER independent of activity states.¹⁰ The relationship of ARI to symptoms in different vigilance states is not clear in neonates presenting with brief resolved unexplained events (BRUE), and the necessity to treat the intermediate ARI group has been uncertain. Therefore, our objective was to describe the pattern of GER events in wake and sleep states with increasing ARI in neonates, and to test the hypothesis that GER related symptoms are frequent in ARI> 7% in wake-state.

Methods

Twenty-five infants, referred for the investigation of suspected GER and BRUE, underwent concurrent pH-impedance and video-polysomnography (PSG) at The Neonatal and Infant Feeding Disorder Program and Sleep Disorder Center at Nationwide Children’s Hospital. Presenting troublesome symptoms to parents and providers included life-threatening events (5), apnea (14), desaturations (10), bradycardia (11), cyanosis (8); individually or in combination. Informed consent was obtained from parent(s) and the study protocol was approved by the ethics committee at the Institutional Research Review Board at the Nationwide Children's Hospital Research Institute, Columbus, OH.

The pH-impedance methods have been described in detail by our group.^10,12 Briefly, studies were performed using ambulatory pH-impedance recorder (Ohmega, Medical Measurement Systems [MMS] Inc. Dover, NH, and USA). pH-Impedance catheters (MMS), 6.4 French, with six impedance channels and one pH channel were used. The catheter was first calibrated and then passed nasally so that the pH electrode was positioned at 87% of the distance from nares to the upper border of lower esophageal sphincter (LES) and position was confirmed by X-ray.¹³

Approximately six hours of bedside polysomnography concurrent with pH-impedance studies was performed according to American Academy of Sleep Medicine (AASM) standards,^{3, 13} using the Grass Sleep system (Astro-Med; Grass Technologies, West Warwick, RI) and Twin PSG software. All pH-Impedance and polysomnography data were concurrent and synchronized.

Symptoms such as movement, arching, grunting and cardiorespiratory events were visually recorded by trained patient care assistants blinded to the pH-impedance and sleep studies. Certified Sleep Lab Technicians, blinded to GER events, documented apneas, bradycardia, desaturations, hypopnea and periodic breathing on PSG.

pH-Impedance Analyses

pH-Impedance data was analyzed using MMS Software (V.8.21.) and verified by trained researchers. Acid reflux index (ARI) was defined as the percentage of the entire pH record time that esophageal pH was <4.¹ GER events were classified as follows: weakly acidic impedance events, when impedance dropped 50% or more from baseline in two or more consecutive impedance channels and pH was 4 and 7; acidic impedance events, when impedance dropped 50% or more from baseline in two or more consecutive impedance channels and pH was 4; acidic non-impedance events, when pH was <4 and not accompanied by a drop in impedance.¹⁵ Proximal migration was defined as an impedance drop ≥ 50% in the most proximal channel (Z1). Acid clearance time (ACT) was defined as the time taken for the pH to normalize to ≥ 4 for ≥ 5 seconds.¹⁰

Polysomnography Analyses

Polysomnography was scored in 30 sec epochs using standard criteria defined by AASM.^14,15,16 Sleep was classified into active sleep (rapid eye movement, muscular atonia, irregular respiration and heart rate and continuous electroencephalography [EEG] pattern) and quiet sleep (non-rapid eye movement, axial muscle tone, regular respiration and heart rate, and specific EEG pattern). Wake-state was characterized by irregular heart, respiratory activity with muscular activity and changes in EEG pattern. Apnea was defined as interruption of air flow lasting for ≥ 2 breaths. Hypopnea was scored as 50% decrease in airflow or respiratory effort lasting ≥ 2 missed breaths from the end of last normal breathing amplitude associated with ≥ 3% oxygen desaturation. Desaturation was defined as a drop in oxygen saturation ≥ 3% from baseline. Sleep efficiency was the total sleep time that was effectively spent in sleep in bed¹⁷ and apnea-hypopnea index is the number of episodes of apnea and hypopnea per hour of total sleep time.

Symptom Analyses

Symptoms were positively associated with GER event if the GER event occurred within two minutes before the onset of symptoms. Symptom index (SI) was calculated as # of symptoms associated with GER/total # of symptoms in that category*100,¹⁸ with the limitation that increased GER increases the likelihood that a symptom could be associated with reflux by chance..¹⁹ Symptom sensitivity index (SSI) was calculated by formula (# of GER events associated with symptoms/total GER events in that category*100).²⁰ Positive SSI (≥ 10%) when associated with a lower SI indicates that infant’s esophagus is sensitive to reflux and non GER etiology of symptoms should be considered.^19,21,22

Statistical Analyses

Data that were normally distributed were analyzed using unpaired t-tests and data that were not normally distributed were analyzed using Wilcoxon Rank-sum. Values are reported as mean ± standard deviation or median (range) unless otherwise specified. A p-value < 0.05 was considered significant. Infants were categorized into normal (ARI< 3%), intermediate (ARI 3≥ to ≤7%) and abnormal (ARI> 7%) groups.¹ During the preliminary analysis, due to the comparable means between normal and intermediate groups (Figure 1), we combined the data from infants in these two cluster groups into ARI ≤ 7%, and then compared with ARI >7% groups. Data were analyzed using GraphPad PRISM 6 software.

Figure 1.

Comparison of number of reflux events, number of symptoms, ACT and SI between ARI <3% vs. ARI=3–7% groups. All p values are ≥ 0.05 signifying lack of significant difference between means of both groups.

Results

A total of 621 GER events during 163 hours of sleep study were analyzed. Infants were initially grouped into ARI<3% (n=10 subjects, median ARI 1.4 [0–2.8]), ARI 3≥ to ≤7% (n=6 subjects, median ARI 5.3 [4.1–6]) and ARI >7% (n=9 subjects, median ARI 16.4 [7.8–23]).

Three subjects had IVH (33.3%) in the ARI>7 group. BPD was present in 4(40%), 2(33.3%) and 3(33.3%) patients from the groups with ARI <3, 3≥ to ≤7 and >7 respectively (p=1.0). Seven (70%) of those with ARI<3, 3(50%) in the ARI ≥3 to <7 group and 4(44%) in ARI>7 group were born preterm (p=0.6).

In subjects with ARI≤7% vs. ARI>7%, gestational age (34.0 [23.0 – 41.0] vs. 37.0 [30.0 – 40.0] weeks, p = 0.3), birth weight (2.4 [0.5 – 4.1] vs. 2.8 [3.1 – 4.2] kg, p = 0.6), postmenstrual age (42.0 [37.0 – 47.0] vs. 43.0 [44.0 – 50.0] weeks, p = 0.9) and study weight (3.8 [2.4 – 7.2] vs. 3.5 [3.8 – 4.6] kg, p = 0.2) were similar. In ARI≤7%, 14.0 (88.0%) were on full oral feeds at the time of study and 11.0 (69.0%) were breathing room air. In ARI>7%, 6.0 (67.0%) were on full oral feeds at the time of study and 6.0 (67.0%) were breathing room air.

The duration of sleep (244.5(192–277), 255(235–257) and 246(173–297) minutes for ARI <3, 3–7 and >7 respectively; p=1.0) and wake (154.5(96–290), 134(126–145.5), 122.5(73–182) for ARI <3, 3≥ to ≤7and >7; p=0.7 respectively) was not significantly different between groups.

Sleep measurement comparison

There was no difference in sleep measurements between normal (ARI <3%), intermediate (ARI 3≥ to ≤7%) and abnormal (ARI > 7%) groups clustered (Table I). We found no significant linear relationship between sleep measurements and increasing ARI severity (total wake time, slope −2.4 ± 2.1, p=0.3; total sleep time, slope −0.1 ± 1.1, p=0.9; total active sleep, slope −0.2 ± 1.0, p=0.8; total quiet sleep, slope 0.4 ± 0.6, p=0.5; sleep efficiency, slope 0.2 ± 0.3, p=0.5). Similarly, there was no difference in apnea/hypopnea index between different ARI group comparisons (p> 0. 3).

Table 1.

Sleep measurement comparison between ARI < 3%, ARI 3–7 % and ARI > 7% groups

ARI	< 3%	3≥ to ≤7%	> 7%	p-value
Total sleep time, minutes	233.1 ± 14.4	247.2 ± 6.4	246.0 ± 13.0	0.7
Total wake time, minutes	181.2 ± 33.8	135.2 ±7.5	126.4 ± 11.6	0.2
Sleep Efficiency, %	58.7 ± 5.3	64.7 ± 1.4	66.0 ± 3.1	0.4
Total active sleep, minutes	104.4 ± 11.0	121.0 ± 13.8	114.0 ± 12.0	0.6
Total quiet sleep, minutes	117.6 ± 4.3	119.0 ± 9.1	123.6 ± 8.2	0.8

Data are presented as mean ± standard error.

All p values are >0.05 suggesting a lack of difference in sleep measurements between normal, intermediate and abnormal groups.

Pattern of GER events with increasing ARI

During wake-state, increases in ARI were associated with an increase in frequency/hour of acidic non-impedance events (slope 0.2 ± 0.1, p=0.01), acidic impedance events (slope 0.1 ± 0.02, p< 0.0001), ACT for acidic non-impedance events (slope 1.9 ± 0.7, p=0.02) and symptom index (slope 1.8 ± 0.3, p< 0.0001). Similarly, during sleep-state, an increase in ARI was associated with an increase in frequency/hour of weakly acidic impedance events (slope 0.1 ± 0.02, p=0.01), acidic impedance events (slope 0.1 ± 0.01, p=0.0005) and symptom index (slope 1.2 ±0.5, p=0.02).

Comparing between the sleep vs. wake states, significant differences were noted for the frequency of acidic impedance events, acidic-non-impedance events, and symptom sensitivity index, but no differences were noted for symptom index (Figure 2).

Figure 2.

Pattern of frequency of GER events and symptoms with increasing ARI. (* p≤ 0.05). Significant incremental relationship between frequency/hour of A, acidic impedance events with ARI in wake-state (slope 0.1 ± 0.02, p< 0.0001) and sleep-state (slope 0.04 ± 0.01, p=0.0005) and there is a significant difference between the two slopes (p=0.002), B, nonimpedance events with ARI in wake-state (slope 0.2 ± 0.06, p< 0.01) but not in sleep-state (slope 0.06 ± 0.04, p=0.1). There is a significant difference between the two slopes (p=0.03). C, Relationship of symptom sensitivity index with increasing ARI in wake-state (slope −0.2 ± 0.7, p=0.7) and sleep-state (slope 1.4 ± 0.9, p=0.1). There is a significant difference between the two slopes (p=0.04). D, Significant incremental relationship of symptom index with increasing ARI in wake-state (slope 1.8 ± 0.3, p< 0.0001) and sleep-state (slope 1.2 ± 0.5, p=0.02). There is no significant difference between wake-state and sleep-state (p=0.3)

GER characteristics comparison between ARI ≤7% vs. ARI>7%

For ARI ≤7% vs. ARI>7%, the frequency, characteristics of GER events and their symptom associations in wake-state and sleep-state are summarized (Table II). Overall, the frequency of acidic events, proximal migration, acid clearance time and symptom index was higher in ARI>7% vs. ARI≤ 7% in wake and sleep-state.

Table 2.

GER and symptom characteristic between ARI ≤ 7% and >7% groups

Outcomes	ARI ≤ 7%	ARI >7%	P value
Total no. of reflux events
- Wake	11.5 (5.5– 15.0)	24.0 (15.5– 27.0)	0.004*
- Sleep	5.2 ± 3.7	17.1 ± 11.3	0.0007*
Weakly acidic impedance events (#/hr)
- Wake	2.8 ± 2.1	4.5 ± 2.3	0.1
- Sleep	0.6 (0.1– 1.2)	0.9 (0.8– 1.7)	0.1
Acidic impedance events (#/hr)
- Wake	0.0 (0.0– 0.5)	1.9 (1.0– 2.1)	0.0003*
- Sleep	0.0 (0.0– 0.2)	0.5 (0.1– 1.3)	0.005*
Acid clearance time for acidic impedance events
- Wake	49.5 (7.1– 109.5)	185.0 (113.2– 580.4)	0.003*
- Sleep	157.9 (2.8– 273.0)	424.0 (8.5– 653.6)	0.3
Total no. of symptoms
- Wake	26.6 ± 16.6	18.1 ± 10.9	0.2
- Sleep	29.5 (19.5– 42.0)	35.0 (17.5– 67.0)	0.9
Symptom index
- Wake	19.7 ± 8.7	46.9 ± 9.8	<0.0001*
- Sleep	3.2 (1.7– 11.1)	15.9 (2.6– 48.1)	0.1
Symptom sensitivity index
- Wake	42.9 (37.5– 60.0)	31.3 (26.8– 41.8)	0.3
- Sleep	49.7 ± 30.4	39.7 ± 30.2	0.4

Data are presented as mean ± standard error or median (range).

^*p value ≤ 0.05. Frequency of GER events and symptom index are greater in abnormal group with confirmed GER.

When comparing different types of sleep in ARI>7% vs. ≤ 7%, the frequency of acidic impedance events (2.0 [0.0–3.5] vs. 0.0 [0.0–1.0], p=0.04) and acidic non-impedance events (2.5 [0.3–8.0] vs. 0.0 [0.0–1.5], p=0.04) was higher in active sleep. There was no difference between ARI>7% vs. ≤ 7% in quiet sleep for acidic impedance (1.0 [0.0–2.0] vs. 0.0 [0.0–0.5], p=0.1) and acidic non-impedance events (2.0 [0.0–6.5] vs. 0.5 [0.0–2.0], p=0.3). There were no differences in proximal migration or acid clearance time for impedance events when comparing active sleep and quiet sleep in infants with ARI≤ 7% vs. ARI> 7%.

Comparison between intermediate vs. abnormal groups

In order to further establish the difference between the intermediate group (ARI 3≥ to ≤7%), with confirmed GER (ARI>7%), we did additional subgroup analysis. Comparing ARI 3≥ to ≤7% vs. ARI >7%, the frequency/hr of acidic impedance events was lower in wake-state (0.4± 0.6 vs. 1.8 ± 1.3, p=0.03) but not in sleep-state (p=0.1). The frequency/hr of weakly acidic impedance events with significant proximal migration was higher in wake-state for ARI>7% vs. ARI 3≥ to ≤7 %(3.1 ± 1.9 vs. 0.8 ± 1.2, p=0.03), but not in sleep-state (p=0.2). The acid clearance time was higher for all acidic events in wake-state in ARI >7% vs. ARI 3≥ to ≤7 % groups (70.7 [56.1–175.2] vs. 28.4 [17.5–52.6], p=0.007], unlike in sleep-state (p=0.6). The symptom index for all symptoms was higher in wake-state in ARI>7% vs. ARI 3≥ to ≤7 % (46.9 ± 1.0 vs. 21.9 ± 10.1, p=0.0004), but not in sleep-state (p=0.1) (Table III; available at www.jpeds.com). Furthermore, a sub-group comparison was performed within the severe category to test if the sleep outcome measurements between those infants with ARI ≥7 but <10 (N=3) vs. those with ARI ≥10 (N=6) are different. Given the limitations of a small sample size, all p values are >0.05 for total sleep time, total wake time, sleep efficiency, total active sleep time, and total quiet sleep time.

Discussion

Symptoms related to GER and BRUE, in relation to sleep, can be worrisome to parents and providers,¹⁶ and the diagnosis and treatment of such disorders that affect feeding can be challenging.²³ We performed this investigation because such scenarios can cause prolonged hospital stay and re-admissions and delay in feeding progression. Although prior studies have evaluated the relationship between sleep-wake states and GER characteristics in different ARI groups²⁴, the current study assessed for a relationship between the two and compared multiple ARI groups to help develop thresholds for optimal management protocols.

In our study there was no difference in sleep/wake characteristics with a unit-change in ARI. Increasing ARI was associated with increased frequency of acidic events and acid clearance time in wake-state, which was greater than sleep-state. The total number of reflux events and frequency/hour of acidic events were higher in ARI>7% vs. ARI≤ 7% regardless of sleep-wake state. The symptom index was higher in ARI>7 % vs. ARI≤ 7% in wake-state but not in sleep-state. There was no difference in total number of reflux events, total number of symptoms, acid clearance time and symptom index between normal and intermediate groups (ARI <3% vs. ARI 3≥ to ≤7 %) in wake-state and sleep-state. The number of reflux events, acid clearance time and symptom index was more in abnormal group when compared with intermediate group (ARI>7% vs. ARI 3≥ to ≤7 %) in wake-state but not in sleep-state.

In our study, duration of wake-state and sleep-state was similar in different ARI groups which is in agreement with the findings of Ammari et.al.²⁴ Previous studies rationalized that a significant change in sleep measurements was not found because the pH-based reflux index never exceeded 10%¹. However the present study included infants with ARI up to 23%, and found a significant increase in sleep measurements with increasing ARI and across different ARI groups. Although Harris et al found a greater amount of active sleep with higher esophageal acid exposure,²⁵ our cohort showed a non-significant decrease in active sleep with increase in ARI. Further, it is believed that GER frequency is associated with a higher change in sleep stage.²⁴ This cannot be proved unless the change in sleep-state is synchronized with a preceding GER episode to see the frequency of such events. The higher frequency of GER events associated with wake-state does not prove a causal relationship between the two. In the sub-group comparison, we found no differences in sleep efficiency between those infants with ARI ≥7 but <10 vs. those with ARI ≥10, but acknowledge the limitations of a small sample size.

Our group previously reported more GER events during wake-state²⁶ and attributed it to a higher frequency of transient lower esophageal sphincter relaxation and postprandial gastric distention²⁷ during wake-state. However during our analysis of pH studies, meal periods were appropriately excluded. The association of increased acidic event frequency with increased ARI during wake-state indicates a higher sensitivity to acid compared with sleep-state. Furthermore, acidic events were not significantly higher in wake-state when compared with sleep-state with an increase in ARI. GER and symptom causative mechanisms may be enhanced by the sensitization of esophageal afferent neuronal pathways by acidic refluxate.²⁸

The magnitude of ARI (the percentage of time the pH is <4) depends on various factors including frequency of GER events, as well as physical and chemical characteristics and clearance mechanisms of esophageal refluxate. In the current study, higher frequency of acidic impedance events likely contributed to higher esophageal acidification independent of sleep-wake state. Mechanosensitive esophageal mucosal afferent fibers are sensitive to pH.²⁹ Increased acidity in stomach probably activates of mechanoreceptors that trigger transient lower esophageal sphincter relaxation. This agrees with our finding of similar frequency of weakly acidic events in ARI >7 % vs. ARI ≤ 7% groups, regardless of sleep-wake state.

The higher symptom index in wake-state can be attributed to the higher frequency of GER events compared to sleep-state. Therefore, our current findings establish that severity of ARI does not affect symptoms during sleep-state, and that non-GER causes may produce these symptoms. This is in agreement with our previous study showing a low symptom index and high symptom sensitivity index for cardio-respiratory symptoms during sleep-state vs. wake state.²⁶ In our cohort, the symptom indices were higher in the ARI>7% group for somatic symptoms but not for cardio-respiratory or sensory symptoms. Further, the lack of significant differences in symptom sensitivity indices for cardio-respiratory, somatic and sensory symptoms in our study population signifies that life threatening symptoms in infants require further scrutiny and should not be limited to evaluation for GER alone. Further research is needed to understand the role of swallowing, airway-digestive interactions and protective reflexes.

Earlier studies found a higher duration of reflux events during sleep, with quiet sleep durations more prolonged than active sleep.²⁴ This was probably due to marked depression of swallow peristalsis during quiet sleep.²⁴ Consistent with this, regardless of ARI (≤ 7% vs. >7 %), acid clearance time is 2.5 to 3 times greater in sleep-state vs. wake-state. However, using pooled data, we found that for each unit increase in ARI, the rate of change in acid clearance time was less in sleep-state vs. wake-state. This may suggest that in higher ARI states (eg, severe acid GER or esophagitis), sleep-state may rapidly neutralize acid, implying higher sensitization, or increased frequency of swallowing or esophageal peristalsis. We speculate that when such protective aerodigestive mechanisms are not operational, life-threatening symptoms may result.³⁰

Treatment of infants in the intermediate ARI group (between 3–7%) has been controversial. Even though current practice is to treat based on clinical presentation and severity of symptoms, we strongly believe that treatment is unnecessary unless the ARI is > 7%, because long term acid suppression has been associated with side effects³¹, and may affect weight gain, increase risk of infections and prolong hospital stay.² To evaluate this, we compared the GER characteristics in the intermediate group, and found no significant difference vs. the ARI <3% group, and a significant decrease in frequency events and symptom development in wake-state vs. the ARI>7% group. Earlier studies have reported an unlikely association between reflux events and non-esophageal symptoms^32,33 and have discouraged treatment of apnea in premature babies with anti-suppressive medication.³⁴ Our study emphasizes that symptoms during sleep-state are likely due to non-GER related causes, and this helps us reconsider treatment of infants in intermediate group with acid suppressive medication.

In conclusion, a greater magnitude of 24-hr esophageal acidification is associated with increased frequency of reflux in wake-state only. Sleep and supine states are more protective of GER when compared with wake-state, likely due to an elevated chemosensory threshold to acid. It may be sufficient to consider acid suppressive therapy only for infants with ARI> 7% when associated with increased somatic and cardio-respiratory symptoms.

Abbreviations

GER

Gastro-esophageal Reflux

ARI

Acid Reflux Index

BRUE

Brief resolved unexplained events

PSG

Video-Polysomnography

LES

Lower Esophageal Sphincter

ACT

Acid Clearance Time

SI

Symtpom Index

SSI

Symptom Sensitivity Index