- Email: [email protected]
Are Infants with Bronchopulmonary Dysplasia Prone to Gastroesophageal Reflux? A Prospective Observational Study with Esophageal pH-Impedance Monitoring
Are Infants with Bronchopulmonary Dysplasia Prone to Gastroesophageal Reflux? A Prospective Observational Study with Esophageal pH-Impedance Monitoring Stefano Nobile, MD, PhD, Carmine Noviello, MD, PhD, Giovanni Cobellis, MD, PhD, and Virgilio Paolo Carnielli, MD, PhD Objective To perform an observational cohort study with esophageal pH-multichannel intraluminal impedance (pH-MII) monitoring in symptomatic preterm infants with and without bronchopulmonary dysplasia (BPD). Study design We prospectively studied 46 infants born #32 weeks gestational age: 12 infants with BPD and 34 infants without BPD. Each patient had symptoms consistent with gastroesophageal reflux and had 24-hour pH-MII, which were compared between BPD and non-BPD by univariate analysis and quantile regression analysis. Results Demographic and clinical characteristics were similar between infants with and without BPD, except for fluid administration (145 vs 163 mL/kg/d, P = .003), length of stay (92 vs 69 days, P = .019), and time to achieve complete oral feeding (76 vs 51 days, P = .013). The analysis of 1104 hours of pH-MII tracings demonstrated that infants with BPD compared with infants without BPD had increased numbers of pH-only events (median number 21 vs 9) and a higher symptom sensitivity index for pH-only events (9% vs 4.9%); the number and characteristics of acid, weakly acid, nonacid and gas gastroesophageal reflux events, acid exposure, esophageal clearance, and recorded symptoms did not significantly differ between the 2 groups. Conclusions The increased number of (and sensitivity for) pH-only events among infants with BPD may be explained by several factors, including lower milk intake, impaired esophageal motility, and a peculiar autonomic nervous system response pattern. (J Pediatr 2015;-:---).
S
ymptomatic gastroesophageal reflux (GER) has been reported as a common condition among preterm infants and can influence growth, length of hospital stay, and respiratory outcome.1 Bronchopulmonary dysplasia (BPD), characterized by persistent airway inflammation and/or impaired lung alveolarization and growth, is a major complication of prematurity, affecting 30% of extremely low birth weight infants.2,3 Infants with BPD may have an increased risk of GER because of respiratory effort and transient increases in intra-abdominal pressure related to coughing, crying, and air flow obstruction, which can lead to a decrease in lower esophageal sphincter (LES) tone and an increased occurrence of transient LES relaxations.4-6 The use of respiratory stimulants (caffeine) may increase gastric acid secretion, decrease LES pressure, and exacerbate GER.7,8 The role of GER in the pathogenesis and recovery from BPD is controversial.1,6,9,10 Infants with BPD are thought to have a predisposition for GER as some studies reported a high prevalence of GER among infants with BPD, possibly because of microaspirations or reflex mechanisms. Improvements in the respiratory status of infants with BPD have been reported after medical or even surgical antireflux therapy is instituted.5,11-15 However, there are increasing concerns about the safety and efficacy of antireflux drugs among preterm infants, which are increasingly used to treat presumed (symptomatic) but not proven GER.16 Combined esophageal pH-multichannel intraluminal impedance (pH-MII) has emerged as the state-of-the-art method to diagnose GER.17 This method detects and characterizes the retrograde esophageal flow and its subsequent anterograde clearance into the stomach. Therefore, the height, duration, and clearance of refluxate can be reliably characterized to define the presence and pattern of GER events and to find out associations between symptoms and reflux episodes. Given the unclear predisposition for GER among infants with BPD and the few studies about this topic, we used pH-MII to test if symptomatic infants with BPD had more GER than symptomatic infants without BPD.
BPD CPAP GA GER LES MII
Bronchopulmonary dysplasia Continuous positive airway pressure Gestational age Gastroesophageal reflux Lower esophageal sphincter Multichannel intraluminal impedance
NICU pH-MII PMA SAP SI SSI
Neonatal intensive care unit pH-multichannel intraluminal impedance Postmenstrual age Symptom-associated probability Symptom index Symptom sensitivity index
From the Maternal and Child Department, Salesi Children’s Hospital, Ancona, Italy The authors declare no conflicts of interest. Portions of this study were presented as a poster and oral presentation at the European Academy of Paediatric Societies’ meeting, Barcelona, Spain, October 1721, 2014. 0022-3476/$ - see front matter. Copyright ª 2015 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.jpeds.2015.05.005
1
THE JOURNAL OF PEDIATRICS
www.jpeds.com
Methods This observational cohort study was conducted with infants born #32 weeks gestational age (GA) between March 2010 and March 2014 at a regional tertiary-care neonatal intensive care unit (NICU). Inclusion criteria were GA #32 weeks at birth and clinical symptoms suspected to be GER-related (bradycardia, apnea, desaturations, and regurgitations). Bradycardia was defined as heart rate #80 bpm, desaturation as peripheral oxygen saturation #80% (measured by pulse oximeter), and apnea as cessation of breathing for at least 20 seconds (5 seconds if there was a concurrent desaturation).18 Infants were considered for GER evaluation if they presented at least 5 symptoms per day (apnea, desaturations, bradycardia, and regurgitations close to feeds as documented by nursing charts and confirmed by the attending physician) and whose symptoms were not due to other causes (eg, sepsis, seizures, and swallowing incoordination, which were ruled out after careful clinical and laboratory evaluations). Exclusion criteria were major malformations, death or discharge to other hospitals, need for any ventilatory support (ie, nasal continuous positive airway pressure [CPAP], or endotracheal ventilation) at the time of study, and lack of parental consent. A total of 395 infants #32 weeks GA were born in the study period; 9 were excluded for major malformations, 23 died before 36 weeks postmenstrual age (PMA), and 144 were transferred to other hospitals before 36 weeks PMA. Of the remaining 219 patients, 44 (12.1% of live infants) were diagnosed with BPD according to Walsh et al if they failed an oxygen reduction test at 36 (1) weeks PMA.19 Infants who had clinical symptoms consistent with GER and whose parents agreed to the pH-MII testing were studied: 12 out of 44 infants with BPD (10 did not have symptoms, 14 were ventilated or on nasal CPAP, and 8 lacked consent) and 34 out of 175 without BPD (108 did not have symptoms, and 33 lacked consent). No surgeries were performed for GER. Local Institutional Review Board approved the study. pH-MII Methods Each patient had 24-hour pH-MII performed with an esophageal probe connected with a pH-MII device (Ohmega MMS, Enschede, The Netherlands). During the study, all infants were breathing spontaneously, were receiving intermittent bottle or orogastric feeds, and were not on antireflux or antacid medications, inhaled medications, or diuretics. Infants were kept supine or in lateral decubitus during pH-MII. The neonatal pH-MII probe (pHTip EI0634; Unisensor, Attikon, Switzerland) had 6 impedance channels (Z1-Z6) (7 impedance rings positioned 1.5 cm apart from each other) and 1 pH channel on the tip of the catheter. Prior to each study, the pH-MII probe was calibrated in standard pH buffer solutions of 4.0 and 7.0. The probe was then inserted nasally into the esophagus, and the position verified by chest radiograph. Proper location 2
Vol. -, No. of the probe was identified as 2-3 cm above the left diaphragm.20-22 The pH-MII probe was securely taped and connected to the Ohmega system; constant observation and review of the position at the end of the study were assured. pH-MII data were evaluated using manufacturer’s software, and each tracing was manually reviewed and validated by the same operator (S.N.). An acid reflux event was defined as a drop in pH <4 lasting for more than 5 seconds. Liquid reflux was evident when there was a fall in impedance greater than 50% from baseline in at least 2 consecutive channels in an aboral direction. Gasonly reflux was evident as a simultaneous increase in impedance >3000 U in any consecutive impedance sites, with 1 site having an absolute value >7000 U. Mixed reflux events met both criteria.23 On the basis of the lowest pH value recorded during each event, reflux events were divided into acid (pH <4, or occurring when the pH was already acidic), weakly acid (pH between 4 and 7), or nonacid (pH >7). All pH drops <4 exceeding 5 seconds, excluding meal periods, registered independently from the impedance reflux events, were considered pH-only events. Symptoms were documented in real time by NICU nurses who were blinded to the pH-MII recordings. Nurses recorded the beginning and end of each meal and the occurrence of symptoms. An association between symptoms and GER events was considered positive if a symptom occurred within 120 seconds before or after a reflux event.24 For each study, the following variables were evaluated in accordance with the North American Society for Pediatric Gastroenterology, Hepatology and Nutrition/European Society for Paediatric Gastroenterology, Hepatology and Nutrition GER guidelines and with other studies6,21,25: number and pH of reflux episodes, duration of the longest reflux event, number of reflux events lasting >5 minutes, lowest pH, esophageal clearance, acid exposure (area under curve pH <4 in minutes), pHonly events, symptom index (SI), symptom sensitivity index (SSI), and symptom-associated probability (SAP) for liquid and gas refluxes. The SI was calculated as (number of symptoms associated with reflux/number of all symptoms) 100; the SSI was calculated as (number of refluxes associated with symptoms/total number of refluxes in 24 hours) 100; the SAP was calculated for acid, nonacid, and gas reflux using MMS software (Medical Measurement Systems, Enschede, the Netherlands) by dividing 24-hour pH data into consecutive 2-minute segments. We determined for each segment whether reflux and/or symptoms occurred. A 2 2 contingency table was then constructed in which the numbers of segments with and without symptoms and with and without reflux were tabulated. A c2 test was used to calculate the probability (P) that the observed distribution could have been by chance. SAP was calculated as (1 P) 100% and was considered to be positive when $95%.26,27 The pH-MII study was defined as abnormal if the SI was $50%, or the SAP was $95%.25 The number of reflux events reaching the proximal catheters Z1 and Z3, the number of pH-only events, and the associated number of symptoms (within 120 seconds before or after these events) were Nobile et al
- 2015
ORIGINAL ARTICLES
evaluated to describe the spatial extension of reflux. The SI and SSI for reflux events reaching the more proximal impedance channels (Z1 and Z3) and for pH-only events were calculated and compared between infants with and without BPD. An illustrative pH-MII recording is shown in the Figure (available at www.jpeds.com).
weight (1088 vs 1164 g, P = .520), age at evaluation (48 vs 29 days, corresponding to 36.1 vs 33.4 weeks PMA, P = .209), weight at evaluation (1896 vs 1810 g, P = .694), recorded symptoms during pH-MII (median number 6 vs 12, P = .456), caffeine therapy at the time of pH-MII (33.3 vs 47.0%, P = .509), postnatal steroid administration (25% vs 2.9%, P = .049), intraventricular hemorrhage grade 3 or higher (0% vs 2.9%, P = 1.000), periventricular leukomalacia (0% vs 8.8%, P = .557), seizures (0% vs 2.9%, P = 1.000), hematocrit at study (mean 33.8% vs 33.1%, P = .074). Regarding growth and nutrition, the following characteristics were not significantly different between groups: partial or complete orogastric tube feeding (76.5% vs 83.3%, P = 1.000), prevalence of (>50% of the total) human milk feeding (50% vs 52.9%, P = 1.000), use of milk fortifiers (58.3% vs 38.2%, P = .314), weight gain for the preceding 3 days (14.5 vs 13.6 g/kg/d, P = .930) and 7 days (15.8 vs 16.4 g/kg/d, P = .990), energy intake on the day of the study (131.5 vs 131.2 kCal/kg/d, P = .410), and oral lipid intake on the day of the study (6.9 vs 6.8 g/kg/d, P = .726). Fluid administration at the time of study, length of stay, and time to achieve complete oral feeding were significantly different between groups (Table I). In total, 1104 hours of tracings were analyzed (mean 24 h/ subject) and 765 symptoms were recorded (mean 16.6 per patient) during the pH-MII monitoring; 188 out of 765 symptoms (25%) were temporally associated with GER events. Esophageal pH monitoring enabled us to detect and characterize 819 acid reflux events and 661 pH-only events. MII allowed us to demonstrate 1702 GER events (203 acid, 1037 weakly acid, 17 nonacid, and 445 gas). Regarding pHMII variables, there were no significant differences at univariate analysis between infants with and without BPD, even if infants with BPD had higher numbers of reflux events and pH-only events, and infants without BPD had longer GER events (Table II). With regard to the possibility that symptoms such as apnea and desaturations may lead to relaxation of LES and, thus, precede GER events, we found that 83 symptoms (11% of the detected symptoms) preceded GER events: 20 out of 153 (13%) in infants with BPD and 63 out of 612 (10%) in infants without BPD. We then evaluated the proximal extension of reflux episodes and found that 114 out of 1702 (6.7%) reached the most proximal channel Z1 and 513 out of 1702 (30.1%) reached Z3. Of 114 reflux events reaching Z1, 8
Clinical Data Demographic and clinical data were prospectively collected according to predefined criteria, and extracted from our electronic database (NeoTools; Interactive, Milano, Italy). Data are expressed as means (range, IQR), medians (for not normally distributed data), and percentages for categorical data. The t test, Mann-Whitney test, and c2 analysis-Fisher exact test were performed as appropriate to compare variables between BPD and controls using SPSS software (v 19.0; IBM, Armonk, New York). With regards to the comparison of demographic and clinical characteristics, we decreased the alpha level by the number of comparisons made in order to avoid increases in type I error, and the resulting P was .038; the same was done for the pH-MII data, and the resulting P levels were .005 (for pH data) and .0038 (for multichannel intraluminal impedance [MII] data). A linear quantile regression analysis was performed to evaluate the effect of BPD, caffeine treatment, orogastric tube on the total number of reflux events, and also on each type of reflux episode, used as dependent variables. The quantile regression analysis is a nonparametric method, which evaluates the effect of an exposure on different parts of the outcome distribution, estimating a quantile-specific effect. In our analysis, we evaluated the effect of the independent variable on the median value of the total number of reflux events distribution.28 For each model, point regression coefficient estimates and 95% CIs were calculated. When the 95% CI did not include 0, the regression coefficient was considered statistically significant.
Results The study population consisted of 46 preterm infants (25 male, 21 female): 12 of them (8 male) were diagnosed with BPD and 34 (17 male) infants did not have BPD. Demographic and clinical characteristics were similar between infants with and without BPD: male sex (66.7% vs 50.0%, P = .502), GA at birth (29.4 vs 29.1 weeks, P = .639), birth
Table I. Demographic and clinical characteristics of patients Fluid administration during study (mL/kg/d), mean (SD, range) Length of stay (d), mean (SD, range) Time to achieve complete oral feeding (d), median (range)
BPD (n = 12)
No BPD (n = 34)
P
145 (11.5, 131-162) 92 (26, 56-136) 76 (44-122)
163 (10.1, 140-178) 69 (23, 32-126) 51 (28-115)
.003* .019* .013†
*t test. †Mann-Whitney test.
Are Infants with Bronchopulmonary Dysplasia Prone to Gastroesophageal Reflux? A Prospective Observational Study with Esophageal pH-Impedance Monitoring
3
THE JOURNAL OF PEDIATRICS
Vol. -, No. -
www.jpeds.com
Table II. pH-MII comparison between infants with and without BPD (univariate analysis) pH measurements Number of reflux events, median (SD, 95% CI) Reflux index (n of reflux events/h), median (SD, 95% CI) Number of reflux events lasting >5 min, median (SD, 95% CI) Longest reflux event (min), median (SD, 95% CI) Lower pH, mean (SD, 95% CI) Esophageal clearance (min/reflux), median (SD, 95% CI) pH-only events (n), median (SD, 95% CI) Area under the curve pH <4 (min), median (SD, 95% CI) Area under the curve pH <4 lasting $10%, n (%) SI (acid reflux events) $50%, n (%) SAP (acid reflux events) $95%, n (%) MII measurements Number of acid reflux events, median (SD, 95% CI) Number of weakly acid reflux events, median (SD, 95% CI) Number of nonacid reflux events, median (SD, 95% CI) Number of gas reflux events, median (SD, 95% CI) SI (acid reflux events) $50%, n (%) SAP (acid reflux events) $95%, n (%) SI (weakly acid reflux events) $50%, n (%) SAP (weakly acid reflux events) $95%, n (%) SI (gas reflux events) $50%, n (%) SAP (gas reflux events) $95%, n (%) Reflux events reaching Z1-associated symptoms (n), median (SD, 95% CI) Reflux events reaching Z3 (n), median (SD, 95% CI) Reflux events reaching Z3-associated symptoms (n), median (SD, 95% CI) pH-only events-associated symptoms (n), median (SD, 95% CI)
BPD (n = 12)
No BPD (n = 34)
P
24.0 (27.0, 7.6-52.7) 1.3 (1.1, 0.4-2.3) 1.5 (5.4, 0.1-9.1) 8.0 (112.3, 25.6 to 135.1) 1.54 (0.8, 0.5-1.8) 2.9 (3.3, 1.0-6.5) 21.0 (26.5, 6.1-50.4) 27.0 (376.6, 60.9 to 518.1) 3 (25.0%) 0 (0) 1 (8.3%)
11.0 (20.0, 10.5-27.8) 0.5 (0.7, 0.4-1.1) 2.0 (2.6, 1.5-3.7) 14.6 (14.5, 10.2-21.0) 2.08 (1.1, 1.6-2.5) 2.3 (2.8, 2.1-4.5) 9.0 (18.2, 8.0-23.8) 40.5 (58.1, 31.5-81.8) 3 (8.8%) 1 (2.9) 2 (5.9%)
.416† .420† .974† .673† .175* .710† .360† .599† .309z 1.000z 1.000z
3.0 (4.3, 0.5-8.4) 13.0 (18.8, 0.1-35.0) 5.0 (13.7, 1.5-23.8) 1 (8.3) 1 (8.3) 0 (0) 1 (8.3) 0 (0) 2 (16.7) 0 (0.5, 0.2 to 0.7) 3.0 (14.0, 4.0 to 22.0) 1.0 (1.5, 0.1-2.7) 0 (5.6, 2.2 to 8.2)
2.0 (7.6, 2.2-8.8) 14.0 (20.2, 12.5-30.0) 0 (2.8, 0.4 to 1.9) 11.0 (7.2, 8.0-14.3) 1 (2.9) 5 (14.7) 2 (5.9) 5 (14.7) 3 (8.8) 6 (17.6) 0 (0.8, 0.1 to 0.6) 8.0 (12.1, 5.4-15.8) 1.0 (2.1, 0.6-2.4) 0 (2.2, 0-1.9)
.455† .832† .306† .435† .432z 1.000z 1.000z 1.000z .557z 1.000z .525† .694† .985† .526†
*t test. †Mann-Whitney test. zc2 test/Fisher exact test.
(SSI = 7.0%) were associated with symptoms: 2 infants with BPD (SSI = 8.3%) and 10 infants without BPD (SSI = 6.7%); the SSI was quite similar between the groups. Of 513 reflux events reaching Z3, 43 (SSI = 8.4%) were associated with symptoms: 9 infants with BPD (SSI = 8.1%) and 34 infants without BPD (SSI = 8.5%); again, the SSI was comparable between the groups. Regarding pH-only events, 42 out of 661 (SSI = 6.4%) were associated with symptoms: 21 infants with BPD (SSI = 9.0%) and 21 infants without BPD (SSI = 4.9%); the SSI was nearly double in infants with BPD vs those without BPD. Thirteen out of forty-six patients (28%) received medications for GER after pH-MII: 11 omeprazole, 1 omeprazole and domperidone, and 1 ranitidine and domperidone. Eight out of 13 treated patients (62%) responded to the treatment after 8 days (mean), and 5 showed a later (after $14 days) decrease in symptoms that was not attributed to the therapy. All the study patients were discharged home after spending at least 3 days free of desaturations, bradycardia, or apnea. No patients had occurrence of symptoms after the pH-MII study. All infants with BPD were discharged home in room air. Table III shows the results of multiple quantile regression analysis. In particular, orogastric tube feeding was significantly associated with acid reflux events and inversely correlated with gas reflux events. Concomitant caffeine therapy was significantly associated with a lower number of weakly acid reflux episodes. BPD had a significant positive effect on pH-only reflux events. 4
Discussion Infants with BPD did not have significantly higher GER features compared with infants without BPD as measured by esophageal pH-MII monitoring, except for higher occurrence of pH-only events and higher SSI for pH-only events.
Table III. Effect of BPD, caffeine treatment, and orogastric tube feeding on GER events: results of the quantile regression analysis Total number of reflux events BPD (yes vs no) Caffeine (yes vs no) Orogastric tube feeding (yes vs no) Total acid reflux events BPD (yes vs no) Caffeine (yes vs no) Orogastric tube feeding (yes vs no) Total weakly acid reflux events BPD (yes vs no) Caffeine (yes vs no) Orogastric tube feeding (yes vs no) Total gas reflux events BPD (yes vs no) Caffeine (yes vs no) Orogastric tube feeding (yes vs no) Total pH-only events BPD (yes vs no) Caffeine (yes vs no) Orogastric tube feeding (yes vs no)
Coefficient
95% CI
4.25 29.25 42.75
15.25; 47.11 36.82; 17.82 8.11; 70.65
1.50 8.00 11.00
12.25; 7.88 26.17; 17.99 1.49; 26.69
1.89 11.33 3.78
15.19; 4.48 24.08; 3.81 22.02; 11.83
2.15 0.77 6.85
2.18; 11.00 4.79; 5.68 9.73; 0.90
9.00 3.17 5.42
0.22; 18.07 0.41; 12.42 2.26; 23.07
Nobile et al
- 2015 pH-only events are acid reflux events lasting >5 seconds that are detected by pH-metry and not identified by MII. pH-only events may depend on immature peristalsis of the esophagus, lower LES tone, and delayed gastric empting (thus, explaining their increased frequency among preterm infants compared with older children and adults) or to other factors such as reflux esophagitis, insignificant bolus size, or slow pH drift.22,29 We speculate that: (1) infants with BPD might have impaired esophageal motility, decreased basal LES tone, and increased LES relaxations, as pointed out by others5; and (2) patients with BPD could be more sensitive than infants without BPD to the autonomic nervous system responses triggered by pH-only events that result in clinical symptoms. The latter point could be explained by the hypothesis that the esophageal afferent neural pathways might be more reactive to decreases in pH among patients with BPD because of their increased inflammatory state.30 The activation of vagal afferents could trigger airway efferent responses resulting in protective reflexes such as esophagoglottal closure and pharyngoglottal closure.31 Another explanation of our findings could rely on the fact that infants with BPD compared with infants without BPD received a significantly lower intake of fluids at the time of pH-MII because of their chronic pulmonary condition, resulting in a lower buffering capacity of milk. However, the energy intake, the type of milk and the oral lipid amount were similar in the 2 groups. It is unlikely that the feeding modality could have influenced the results because a similar proportion of patients in both groups were fed by an orogastric tube, which is a factor that could have influenced GER.9,10,29,32 The influence of the pH-MII probe itself may theoretically promote reflux; however, it is an intrinsic, unavoidable feature of the pH-MII method. The feeding duration could influence the number and characteristics of GER events.33 The feeding duration followed our general NICU policy: preterm infants received feeds every 3 hours, and each feed was administered over 20-30 minutes, with small variations depending on infants’ tolerance; it is unlikely that significant differences in feeding duration could have occurred and influenced GER features. We observed that 25% of symptoms were temporally associated with GER events. We think that the majority of symptoms in preterm infants could be explained by functional immaturity of the neuronal control of breathing and/or by autonomic system impairment because of prematurity, as pointed out by other authors.34 The temporal relationship between breathing irregularities and GER events is still controversial, as the percentage of GER-linked apnea (evaluated adopting different definitions and timing) varied from 0% to 94% in previously published studies.35-37 However, there are reports that apnea or periodic breathing in infants may lead to relaxation of LES and, thus, precede GER events.38 Although our study was not aimed at evaluating the relationship between apnea and GER, as esophageal manometry (to better evaluate LES pressure) and polysomnography (to characterize breathing irregularities) would have been needed to address this issue, we
ORIGINAL ARTICLES found that 11% of the symptoms preceded GER events. Because we could not discriminate between apnea and other symptoms, we only speculate that <10% of breathing irregularities (apnea, desaturation, and periodic breathing) may trigger LES relaxations and eventually GER episodes occurring within 120 seconds. Jadcherla et al investigated the role of GER among infants with BPD using esophageal pH-MII,14,15 and other researchers used pH-metry or clinical observations.1,9,10 Jadcherla et al also compared the occurrence of GER and symptom patterns between 18 preterm infants with BPD and 12 without BPD.14 In their work, the frequencies of GER events, acid clearance time, and SSI were all greater in infants with BPD vs those without BPD (P < .001). Higher SSI was noted with pH-only events (P < .0001 vs pH-MII events). Prolonged acid clearance times were associated with symptomatic acid reflux events in infants with BPD but not in those without BPD. However, details about respiratory management were not provided and patients did not have comparable GA at birth; these factors might explain the differences with our results. In another report, Jadcherla et al evaluated the relationship between symptoms and the most proximal extent and duration of acid reflux events in 9 premature and term infants with BPD who had combined esophageal pH-MII and manometry.15 Acid reflux events into the pharynx were associated with increased symptom occurrence and delayed clearance. Symptomatic acid reflux events were associated with longer acid clearance times that were longer the greater the extent the GER events. Overall, the esophageal acid exposure time lasted for about 15% of the day, and the distal esophagus was the most frequent site of proximal acid provocation. However, in this study, 1 patient was on nasal CPAP during the study, and 9 out of 9 were fed by a gastric tube; these factors might have influenced the GER variables.9,32 Compared with the results by Jadcherla et al, we observed that the esophageal acid exposure time was far lower (median 27 minutes per day among infants with BPD and 40.5 minutes per day among those without BPD with 210 minutes per day in the aforementioned study). There also were no differences in the SSI for proximal reflux events. An explanation of these findings could rely either on the absence of ventilatory and intensive support required by our patients or on the greater number of subjects we studied. In our study, infants with BPD compared with those without BPD had prolonged length of stay and took longer to achieve complete oral feeding. This can be explained by longer supplementary oxygen and tube feeding requirement rather than by an increased rate of GER as previously reported by others.10 We reported our experience in GER treatment with medications. The present study was not aimed at evaluating the safety and efficacy of drugs, but considering our results and the increasing concerns about the safety of antacid drugs in newborns, we believe pharmacologic treatment for GER should be initiated only after the demonstration of
Are Infants with Bronchopulmonary Dysplasia Prone to Gastroesophageal Reflux? A Prospective Observational Study with Esophageal pH-Impedance Monitoring
5
THE JOURNAL OF PEDIATRICS
www.jpeds.com
pathologic pH-MII monitoring to avoid unnecessary drug therapy, adverse events, and costs. Another finding of our study was that orogastric tube feeding was significantly associated with acid reflux events and inversely correlated with gas reflux events, as previously suggested.9,32 Finally, caffeine administration at the time of study was significantly associated with a lower number of weakly acid reflux episodes. This finding needs to be confirmed, as caffeine has been previously associated with increased gastric acid secretion and decreased LES tone.7,8 Our study has some limitations. First, there was a selection bias in testing only infants who were symptomatic. However, there are ethical difficulties precluding researchers from performing a 24-hour study in asymptomatic preterm infants to estimate GER.39 Second, although infants with BPD did not show predisposition for GER, it is important to consider that both disorders can be influenced by multiple factors (ie, prematurity per se, caffeine administration among others). Third, the study population was relatively small. To detect a difference of 8% for infants without BPD vs 15% for infants with BPD in having symptoms suggestive for GER, as found in this study (SAP $95% for weakly acid and gas reflux events), a sample of 325 infants with BPD and 325 without BPD would be needed with a power of 0.8 and alpha <.05. In conclusion, our study did not find a higher prevalence of GER features in symptomatic infants with BPD compared with symptomatic preterm infants without BPD, except for greater number of pH-only events and higher SSI for pH-only events. n We are sincerely grateful to Richard J. Grand, MD (Director Emeritus, Harvard Medical School, Boston, Massachusetts), for his valuable support and Rosaria Gesuita, PhD (Universita Politecnica delle Marche, Ancona, Italy), for her statistical contribution and insight. Submitted for publication Oct 26, 2014; last revision received Mar 25, 2015; accepted May 5, 2015.
References 1. Akinola E, Rosenkrantz TS, Pappagallo M, McKay K, Hussain N. Gastroesophageal reflux in infants <32 weeks gestational age at birth: lack of relationship to chronic lung disease. Am J Perinatol 2004;21:57-62. 2. Jobe AH. The new bronchopulmonary dysplasia. Curr Opin Pediatr 2011;23:167-72. 3. Farstad T, Bratlid D, Medbø S, Markestad T , Norwegian Extreme Prematurity Study Group. Bronchopulmonary dysplasia—prevalence, severity, and predictive factors in a national cohort of extremely premature infants. Acta Paediatr 2011;100:53-8. 4. Orenstein SR, Orenstein DM. Gastroesophageal reflux and respiratory disease in children. J Pediatr 1988;112:847-58. 5. Omari T, Barnett C, Snel A, Davidson G, Haslam R, Bakewell M, et al. Mechanism of gastroesophageal reflux in premature infants with chronic lung disease. J Pediatr Surg 1999;34:1795-8. 6. Vandenplas Y, Rudolph CD, Di Lorenzo C, Hassall E, Liptak G, Mazur L, et al. Pediatric Gastroesophageal Reflux Clinical Practice Guidelines: Joint Recommendations of the North American Society for Pediatric Gastroenterology, Hepatology, and Nutrition (NASPGHAN) and the European Society for Pediatric Gastroenterology, Hepatology, and Nutrition (ESPGHAN). J Pediatr Gastroenterol Nutr 2009;49:498-547. 6
Vol. -, No. 7. Foster LJ, Trudeau WL, Goldman AL. Bronchodilator effects on gastric acid secretion. JAMA 1979;241:2613-5. 8. Stein MR, Towner TG, Weber RW, Mansfield LE, Jacobson KW, McDonnell JT, et al. The effect of theophylline on the lower esophageal sphincter pressure. Ann Allergy 1980;45:238-41. 9. Sindel B, Maisels MJ, Ballantine TVN. Gastroesophageal reflux to the proximal esophagus in infants with bronchopulmonary dysplasia. Am J Dis Child 1989;143:1103-6. 10. Khalaf MN, Porat R, Brodsky NL, Bhandari V. Clinical correlations in infants in the neonatal intensive care unit with varying severity of gastroesophageal reflux. J Pediatr Gastroenterol Nutr 2001;32:45-9. 11. Giuffre RM, Rubin S, Mitchell I. Antireflux surgery in infants with bronchopulmonary dysplasia. Am J Dis Child 1987;141:648-51. 12. Hoyoux C, Forget P, Lambrechts L, Geubelle F. Chronic bronchopulmonary disease and gastroesophageal reflux in children. Pediatr Pulmonol 1985;3:149-53. 13. St Cyr JA, Ferrara TB, Thompson T, Johnson D, Foker JE. Treatment of pulmonary manifestations of gastroesophageal reflux in children two years of age or less. Am J Surg 1989;157:400-4. 14. Jadcherla SR, Peng J, Chan CY, Moore R, Wei L, Fernandez S, et al. Significance of gastroesophageal refluxate in relation to physical, chemical and spatio-temporal characteristics in symptomatic ICU neonates. Pediatr Res 2011;70:192-8. 15. Jadcherla S, Gupta A, Fernandez S, Nelin LD, Castile R, Gest AL, et al. Spatiotemporal characteristics of acid refluxate and relationship to symptoms in premature and term infants with chronic lung disease. Am J Gastroenterol 2008;103:720-8. 16. Terrin G, Passariello A, De Curtis M, Manguso F, Salvia G, Lega L, et al. Ranitidine is associated with infections, necrotizing enterocolitis, and fatal outcome in newborns. Pediatrics 2012;129:e40-5. 17. Wenzl TG, Silny J, Schenke S, Peschgens T, Heimann G, Skopnik H. Gastroesophageal reflux and respiratory phenomena in infants: status of the intraluminal impedance technique. J Pediatr Gastroenterol Nutr 1999;28:423-8. 18. Di Fiore JM, Arko MK, Miller MJ, Krauss A, Betkerur A, Zadell A, et al. Cardiorespiratory events in preterm infants referred for apnea monitoring studies. Pediatrics 2001;108:1304-8. 19. Walsh MC, Yao Q, Gettner P, Hale E, Collins M, Hensman A, et al. Impact of a physiologic definition on bronchopulmonary dysplasia rates. Pediatrics 2004;114:1305-11. 20. Condino AA, Sondheimer J, Pan Z, Gralla J, Perry D, O’Connor JA. Evaluation of infantile acid and nonacid gastroesophageal reflux using combined pH monitoring and impedance measurement. J Pediatr Gastroenterol Nutr 2006;42:16-21. 21. Wenzl TG, Benninga MA, Loots CM, Salvatore S, Vandenplas Y, Diaz JA, et al. Indications, methodology, and interpretation of combined esophageal impedance-pH monitoring in children: ESPGHAN EURO-PIG standard protocol. J Pediatr Gastroenterol Nutr 2012;55:230-4. 22. Shin MS, Shim JO, Moon JS, Kim HS, Ko JS, Choi JH, et al. ImpedancepH monitoring and conventional pH monitoring are complementary methods to detect association between gastroesophageal reflux and apnea-related symptoms in preterm infants and neonates. J Matern Fetal Neonatal Med 2012;25:2406-10. 23. Shay S, Tutuian R, Sifrim D, Vela M, Wise J, Balaji N, et al. Twenty-four hour ambulatory simultaneous impedance and pH monitoring: a multicenter report of normal values from 60 healthy volunteers. Am J Gastroenterol 2004;99:1037-43. 24. Sifrim D, Mittal R, Fass R, Smout A, Castell D, Tack J, et al. Acidity and volume of the refluxate in the genesis of gastro-oesophageal reflux disease symptoms. Aliment Pharmacol Ther 2007;25:1003-17. 25. Pilic D, Fr€ ohlich T, N€ oh F, Pappas A, Schmidt-Choudhury A, K€ ohler H, et al. Detection of gastroesophageal reflux in children using combined multichannel intraluminal impedance and pH measurement: data from the German Pediatric Impedance Group. J Pediatr 2011;158:650-4. 26. Breumelhof R, Smout AJ. The symptom sensitivity index: a valuable additional parameter in 24-hour esophageal pH recording. Am J Gastroenterol 1991;86:160-4.
Nobile et al
- 2015 27. Weusten BL, Roelofs JM, Akkermans LM, Van Berge-Henegouwen GP, Smout AJ. The symptom-association probability: an improved method for symptom analysis of 24-hour esophageal pH data. Gastroenterology 1994;107:1741-5. 28. Koenker R. Quantile regression. New York: Cambridge University Press; 2005. 29. Di Fiore JM, Arko M, Churbock K, Hibbs AM, Martin RJ. Technical Limitations in detection of gastroesophageal reflux (GER) in neonates. J Pediatr Gastroenterol Nutr 2009;49:177-82. 30. Bhandari V. Postnatal inflammation in the pathogenesis of bronchopulmonary dysplasia. Birth Defects Res A Clin Mol Teratol 2014;100:189-201. 31. Jadcherla SR, Gupta A, Wang M, Coley BD, Fernandez S, Shaker R. Definition and implications of novel pharyngo-glottal reflex in human infants using concurrent manometry ultrasonography. Am J Gastroenterol 2009;104:2572-82. 32. Solana Garcıa MJ, L opez-Herce Cid J, Sanchez Sanchez C. Gastroesophageal reflux in critically ill children: a review. ISRN Gastroenterol 2013; 2013:824320. 33. Jadcherla SR, Chan CY, Moore R, Malkar M, Timan CJ, Valentine CJ. Impact of feeding strategies on the frequency and clearance of acid
ORIGINAL ARTICLES
34.
35. 36.
37.
38. 39.
and nonacid gastroesophageal reflux events in dysphagic neonates. JPEN J Parenter Enteral Nutr 2012;36:449-55. Eichenwald EC, Aina A, Stark AR. Apnea frequently persists beyond term gestation in infants delivered at 24 to 28 weeks. Pediatrics 1997; 100:354-9. Arad-Cohen N, Cohen A, Tirosh E. The relationship between gastroesophageal reflux and apnea in infants. J Pediatr 2000;137:321-6. Mousa H, Woodley FW, Metheney M, Hayes J. Testing the association between gastroesophageal reflux and apnea in infants. J Pediatr Gastroenterol Nutr 2005;41:169-77. Peter CS, Sprodowski N, Bohnhorst B, Silny J, Poets CF. Gastroesophageal reflux and apnea of prematurity. No temporal relationship. Pediatrics 2002;109:8-11. Omari TI. Apnea-associated reduction in lower esophageal sphincter tone in premature infants. J Pediatr 2009;154:374-8. Vandenplas Y, Belli D, Boige N, Bouquet J, Cadranel S, Cezard JP, et al. A standardized protocol for the methodology of esophageal pH monitoring and interpretation of the data for the diagnosis of gastroesophageal reflux. J Pediatr Gastroenterol Nutr 1992;14: 467-71.
Are Infants with Bronchopulmonary Dysplasia Prone to Gastroesophageal Reflux? A Prospective Observational Study with Esophageal pH-Impedance Monitoring
7
THE JOURNAL OF PEDIATRICS
www.jpeds.com
Vol. -, No. -
Figure. A, 24 hour pH-MII recording. B, Detail of a pH-only event that was followed by a symptom.
7.e1
Nobile et al
S
ymptomatic gastroesophageal reflux (GER) has been reported as a common condition among preterm infants and can influence growth, length of hospital stay, and respiratory outcome.1 Bronchopulmonary dysplasia (BPD), characterized by persistent airway inflammation and/or impaired lung alveolarization and growth, is a major complication of prematurity, affecting 30% of extremely low birth weight infants.2,3 Infants with BPD may have an increased risk of GER because of respiratory effort and transient increases in intra-abdominal pressure related to coughing, crying, and air flow obstruction, which can lead to a decrease in lower esophageal sphincter (LES) tone and an increased occurrence of transient LES relaxations.4-6 The use of respiratory stimulants (caffeine) may increase gastric acid secretion, decrease LES pressure, and exacerbate GER.7,8 The role of GER in the pathogenesis and recovery from BPD is controversial.1,6,9,10 Infants with BPD are thought to have a predisposition for GER as some studies reported a high prevalence of GER among infants with BPD, possibly because of microaspirations or reflex mechanisms. Improvements in the respiratory status of infants with BPD have been reported after medical or even surgical antireflux therapy is instituted.5,11-15 However, there are increasing concerns about the safety and efficacy of antireflux drugs among preterm infants, which are increasingly used to treat presumed (symptomatic) but not proven GER.16 Combined esophageal pH-multichannel intraluminal impedance (pH-MII) has emerged as the state-of-the-art method to diagnose GER.17 This method detects and characterizes the retrograde esophageal flow and its subsequent anterograde clearance into the stomach. Therefore, the height, duration, and clearance of refluxate can be reliably characterized to define the presence and pattern of GER events and to find out associations between symptoms and reflux episodes. Given the unclear predisposition for GER among infants with BPD and the few studies about this topic, we used pH-MII to test if symptomatic infants with BPD had more GER than symptomatic infants without BPD.
BPD CPAP GA GER LES MII
Bronchopulmonary dysplasia Continuous positive airway pressure Gestational age Gastroesophageal reflux Lower esophageal sphincter Multichannel intraluminal impedance
NICU pH-MII PMA SAP SI SSI
Neonatal intensive care unit pH-multichannel intraluminal impedance Postmenstrual age Symptom-associated probability Symptom index Symptom sensitivity index
From the Maternal and Child Department, Salesi Children’s Hospital, Ancona, Italy The authors declare no conflicts of interest. Portions of this study were presented as a poster and oral presentation at the European Academy of Paediatric Societies’ meeting, Barcelona, Spain, October 1721, 2014. 0022-3476/$ - see front matter. Copyright ª 2015 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.jpeds.2015.05.005
1
THE JOURNAL OF PEDIATRICS
www.jpeds.com
Methods This observational cohort study was conducted with infants born #32 weeks gestational age (GA) between March 2010 and March 2014 at a regional tertiary-care neonatal intensive care unit (NICU). Inclusion criteria were GA #32 weeks at birth and clinical symptoms suspected to be GER-related (bradycardia, apnea, desaturations, and regurgitations). Bradycardia was defined as heart rate #80 bpm, desaturation as peripheral oxygen saturation #80% (measured by pulse oximeter), and apnea as cessation of breathing for at least 20 seconds (5 seconds if there was a concurrent desaturation).18 Infants were considered for GER evaluation if they presented at least 5 symptoms per day (apnea, desaturations, bradycardia, and regurgitations close to feeds as documented by nursing charts and confirmed by the attending physician) and whose symptoms were not due to other causes (eg, sepsis, seizures, and swallowing incoordination, which were ruled out after careful clinical and laboratory evaluations). Exclusion criteria were major malformations, death or discharge to other hospitals, need for any ventilatory support (ie, nasal continuous positive airway pressure [CPAP], or endotracheal ventilation) at the time of study, and lack of parental consent. A total of 395 infants #32 weeks GA were born in the study period; 9 were excluded for major malformations, 23 died before 36 weeks postmenstrual age (PMA), and 144 were transferred to other hospitals before 36 weeks PMA. Of the remaining 219 patients, 44 (12.1% of live infants) were diagnosed with BPD according to Walsh et al if they failed an oxygen reduction test at 36 (1) weeks PMA.19 Infants who had clinical symptoms consistent with GER and whose parents agreed to the pH-MII testing were studied: 12 out of 44 infants with BPD (10 did not have symptoms, 14 were ventilated or on nasal CPAP, and 8 lacked consent) and 34 out of 175 without BPD (108 did not have symptoms, and 33 lacked consent). No surgeries were performed for GER. Local Institutional Review Board approved the study. pH-MII Methods Each patient had 24-hour pH-MII performed with an esophageal probe connected with a pH-MII device (Ohmega MMS, Enschede, The Netherlands). During the study, all infants were breathing spontaneously, were receiving intermittent bottle or orogastric feeds, and were not on antireflux or antacid medications, inhaled medications, or diuretics. Infants were kept supine or in lateral decubitus during pH-MII. The neonatal pH-MII probe (pHTip EI0634; Unisensor, Attikon, Switzerland) had 6 impedance channels (Z1-Z6) (7 impedance rings positioned 1.5 cm apart from each other) and 1 pH channel on the tip of the catheter. Prior to each study, the pH-MII probe was calibrated in standard pH buffer solutions of 4.0 and 7.0. The probe was then inserted nasally into the esophagus, and the position verified by chest radiograph. Proper location 2
Vol. -, No. of the probe was identified as 2-3 cm above the left diaphragm.20-22 The pH-MII probe was securely taped and connected to the Ohmega system; constant observation and review of the position at the end of the study were assured. pH-MII data were evaluated using manufacturer’s software, and each tracing was manually reviewed and validated by the same operator (S.N.). An acid reflux event was defined as a drop in pH <4 lasting for more than 5 seconds. Liquid reflux was evident when there was a fall in impedance greater than 50% from baseline in at least 2 consecutive channels in an aboral direction. Gasonly reflux was evident as a simultaneous increase in impedance >3000 U in any consecutive impedance sites, with 1 site having an absolute value >7000 U. Mixed reflux events met both criteria.23 On the basis of the lowest pH value recorded during each event, reflux events were divided into acid (pH <4, or occurring when the pH was already acidic), weakly acid (pH between 4 and 7), or nonacid (pH >7). All pH drops <4 exceeding 5 seconds, excluding meal periods, registered independently from the impedance reflux events, were considered pH-only events. Symptoms were documented in real time by NICU nurses who were blinded to the pH-MII recordings. Nurses recorded the beginning and end of each meal and the occurrence of symptoms. An association between symptoms and GER events was considered positive if a symptom occurred within 120 seconds before or after a reflux event.24 For each study, the following variables were evaluated in accordance with the North American Society for Pediatric Gastroenterology, Hepatology and Nutrition/European Society for Paediatric Gastroenterology, Hepatology and Nutrition GER guidelines and with other studies6,21,25: number and pH of reflux episodes, duration of the longest reflux event, number of reflux events lasting >5 minutes, lowest pH, esophageal clearance, acid exposure (area under curve pH <4 in minutes), pHonly events, symptom index (SI), symptom sensitivity index (SSI), and symptom-associated probability (SAP) for liquid and gas refluxes. The SI was calculated as (number of symptoms associated with reflux/number of all symptoms) 100; the SSI was calculated as (number of refluxes associated with symptoms/total number of refluxes in 24 hours) 100; the SAP was calculated for acid, nonacid, and gas reflux using MMS software (Medical Measurement Systems, Enschede, the Netherlands) by dividing 24-hour pH data into consecutive 2-minute segments. We determined for each segment whether reflux and/or symptoms occurred. A 2 2 contingency table was then constructed in which the numbers of segments with and without symptoms and with and without reflux were tabulated. A c2 test was used to calculate the probability (P) that the observed distribution could have been by chance. SAP was calculated as (1 P) 100% and was considered to be positive when $95%.26,27 The pH-MII study was defined as abnormal if the SI was $50%, or the SAP was $95%.25 The number of reflux events reaching the proximal catheters Z1 and Z3, the number of pH-only events, and the associated number of symptoms (within 120 seconds before or after these events) were Nobile et al
- 2015
ORIGINAL ARTICLES
evaluated to describe the spatial extension of reflux. The SI and SSI for reflux events reaching the more proximal impedance channels (Z1 and Z3) and for pH-only events were calculated and compared between infants with and without BPD. An illustrative pH-MII recording is shown in the Figure (available at www.jpeds.com).
weight (1088 vs 1164 g, P = .520), age at evaluation (48 vs 29 days, corresponding to 36.1 vs 33.4 weeks PMA, P = .209), weight at evaluation (1896 vs 1810 g, P = .694), recorded symptoms during pH-MII (median number 6 vs 12, P = .456), caffeine therapy at the time of pH-MII (33.3 vs 47.0%, P = .509), postnatal steroid administration (25% vs 2.9%, P = .049), intraventricular hemorrhage grade 3 or higher (0% vs 2.9%, P = 1.000), periventricular leukomalacia (0% vs 8.8%, P = .557), seizures (0% vs 2.9%, P = 1.000), hematocrit at study (mean 33.8% vs 33.1%, P = .074). Regarding growth and nutrition, the following characteristics were not significantly different between groups: partial or complete orogastric tube feeding (76.5% vs 83.3%, P = 1.000), prevalence of (>50% of the total) human milk feeding (50% vs 52.9%, P = 1.000), use of milk fortifiers (58.3% vs 38.2%, P = .314), weight gain for the preceding 3 days (14.5 vs 13.6 g/kg/d, P = .930) and 7 days (15.8 vs 16.4 g/kg/d, P = .990), energy intake on the day of the study (131.5 vs 131.2 kCal/kg/d, P = .410), and oral lipid intake on the day of the study (6.9 vs 6.8 g/kg/d, P = .726). Fluid administration at the time of study, length of stay, and time to achieve complete oral feeding were significantly different between groups (Table I). In total, 1104 hours of tracings were analyzed (mean 24 h/ subject) and 765 symptoms were recorded (mean 16.6 per patient) during the pH-MII monitoring; 188 out of 765 symptoms (25%) were temporally associated with GER events. Esophageal pH monitoring enabled us to detect and characterize 819 acid reflux events and 661 pH-only events. MII allowed us to demonstrate 1702 GER events (203 acid, 1037 weakly acid, 17 nonacid, and 445 gas). Regarding pHMII variables, there were no significant differences at univariate analysis between infants with and without BPD, even if infants with BPD had higher numbers of reflux events and pH-only events, and infants without BPD had longer GER events (Table II). With regard to the possibility that symptoms such as apnea and desaturations may lead to relaxation of LES and, thus, precede GER events, we found that 83 symptoms (11% of the detected symptoms) preceded GER events: 20 out of 153 (13%) in infants with BPD and 63 out of 612 (10%) in infants without BPD. We then evaluated the proximal extension of reflux episodes and found that 114 out of 1702 (6.7%) reached the most proximal channel Z1 and 513 out of 1702 (30.1%) reached Z3. Of 114 reflux events reaching Z1, 8
Clinical Data Demographic and clinical data were prospectively collected according to predefined criteria, and extracted from our electronic database (NeoTools; Interactive, Milano, Italy). Data are expressed as means (range, IQR), medians (for not normally distributed data), and percentages for categorical data. The t test, Mann-Whitney test, and c2 analysis-Fisher exact test were performed as appropriate to compare variables between BPD and controls using SPSS software (v 19.0; IBM, Armonk, New York). With regards to the comparison of demographic and clinical characteristics, we decreased the alpha level by the number of comparisons made in order to avoid increases in type I error, and the resulting P was .038; the same was done for the pH-MII data, and the resulting P levels were .005 (for pH data) and .0038 (for multichannel intraluminal impedance [MII] data). A linear quantile regression analysis was performed to evaluate the effect of BPD, caffeine treatment, orogastric tube on the total number of reflux events, and also on each type of reflux episode, used as dependent variables. The quantile regression analysis is a nonparametric method, which evaluates the effect of an exposure on different parts of the outcome distribution, estimating a quantile-specific effect. In our analysis, we evaluated the effect of the independent variable on the median value of the total number of reflux events distribution.28 For each model, point regression coefficient estimates and 95% CIs were calculated. When the 95% CI did not include 0, the regression coefficient was considered statistically significant.
Results The study population consisted of 46 preterm infants (25 male, 21 female): 12 of them (8 male) were diagnosed with BPD and 34 (17 male) infants did not have BPD. Demographic and clinical characteristics were similar between infants with and without BPD: male sex (66.7% vs 50.0%, P = .502), GA at birth (29.4 vs 29.1 weeks, P = .639), birth
Table I. Demographic and clinical characteristics of patients Fluid administration during study (mL/kg/d), mean (SD, range) Length of stay (d), mean (SD, range) Time to achieve complete oral feeding (d), median (range)
BPD (n = 12)
No BPD (n = 34)
P
145 (11.5, 131-162) 92 (26, 56-136) 76 (44-122)
163 (10.1, 140-178) 69 (23, 32-126) 51 (28-115)
.003* .019* .013†
*t test. †Mann-Whitney test.
Are Infants with Bronchopulmonary Dysplasia Prone to Gastroesophageal Reflux? A Prospective Observational Study with Esophageal pH-Impedance Monitoring
3
THE JOURNAL OF PEDIATRICS
Vol. -, No. -
www.jpeds.com
Table II. pH-MII comparison between infants with and without BPD (univariate analysis) pH measurements Number of reflux events, median (SD, 95% CI) Reflux index (n of reflux events/h), median (SD, 95% CI) Number of reflux events lasting >5 min, median (SD, 95% CI) Longest reflux event (min), median (SD, 95% CI) Lower pH, mean (SD, 95% CI) Esophageal clearance (min/reflux), median (SD, 95% CI) pH-only events (n), median (SD, 95% CI) Area under the curve pH <4 (min), median (SD, 95% CI) Area under the curve pH <4 lasting $10%, n (%) SI (acid reflux events) $50%, n (%) SAP (acid reflux events) $95%, n (%) MII measurements Number of acid reflux events, median (SD, 95% CI) Number of weakly acid reflux events, median (SD, 95% CI) Number of nonacid reflux events, median (SD, 95% CI) Number of gas reflux events, median (SD, 95% CI) SI (acid reflux events) $50%, n (%) SAP (acid reflux events) $95%, n (%) SI (weakly acid reflux events) $50%, n (%) SAP (weakly acid reflux events) $95%, n (%) SI (gas reflux events) $50%, n (%) SAP (gas reflux events) $95%, n (%) Reflux events reaching Z1-associated symptoms (n), median (SD, 95% CI) Reflux events reaching Z3 (n), median (SD, 95% CI) Reflux events reaching Z3-associated symptoms (n), median (SD, 95% CI) pH-only events-associated symptoms (n), median (SD, 95% CI)
BPD (n = 12)
No BPD (n = 34)
P
24.0 (27.0, 7.6-52.7) 1.3 (1.1, 0.4-2.3) 1.5 (5.4, 0.1-9.1) 8.0 (112.3, 25.6 to 135.1) 1.54 (0.8, 0.5-1.8) 2.9 (3.3, 1.0-6.5) 21.0 (26.5, 6.1-50.4) 27.0 (376.6, 60.9 to 518.1) 3 (25.0%) 0 (0) 1 (8.3%)
11.0 (20.0, 10.5-27.8) 0.5 (0.7, 0.4-1.1) 2.0 (2.6, 1.5-3.7) 14.6 (14.5, 10.2-21.0) 2.08 (1.1, 1.6-2.5) 2.3 (2.8, 2.1-4.5) 9.0 (18.2, 8.0-23.8) 40.5 (58.1, 31.5-81.8) 3 (8.8%) 1 (2.9) 2 (5.9%)
.416† .420† .974† .673† .175* .710† .360† .599† .309z 1.000z 1.000z
3.0 (4.3, 0.5-8.4) 13.0 (18.8, 0.1-35.0) 5.0 (13.7, 1.5-23.8) 1 (8.3) 1 (8.3) 0 (0) 1 (8.3) 0 (0) 2 (16.7) 0 (0.5, 0.2 to 0.7) 3.0 (14.0, 4.0 to 22.0) 1.0 (1.5, 0.1-2.7) 0 (5.6, 2.2 to 8.2)
2.0 (7.6, 2.2-8.8) 14.0 (20.2, 12.5-30.0) 0 (2.8, 0.4 to 1.9) 11.0 (7.2, 8.0-14.3) 1 (2.9) 5 (14.7) 2 (5.9) 5 (14.7) 3 (8.8) 6 (17.6) 0 (0.8, 0.1 to 0.6) 8.0 (12.1, 5.4-15.8) 1.0 (2.1, 0.6-2.4) 0 (2.2, 0-1.9)
.455† .832† .306† .435† .432z 1.000z 1.000z 1.000z .557z 1.000z .525† .694† .985† .526†
*t test. †Mann-Whitney test. zc2 test/Fisher exact test.
(SSI = 7.0%) were associated with symptoms: 2 infants with BPD (SSI = 8.3%) and 10 infants without BPD (SSI = 6.7%); the SSI was quite similar between the groups. Of 513 reflux events reaching Z3, 43 (SSI = 8.4%) were associated with symptoms: 9 infants with BPD (SSI = 8.1%) and 34 infants without BPD (SSI = 8.5%); again, the SSI was comparable between the groups. Regarding pH-only events, 42 out of 661 (SSI = 6.4%) were associated with symptoms: 21 infants with BPD (SSI = 9.0%) and 21 infants without BPD (SSI = 4.9%); the SSI was nearly double in infants with BPD vs those without BPD. Thirteen out of forty-six patients (28%) received medications for GER after pH-MII: 11 omeprazole, 1 omeprazole and domperidone, and 1 ranitidine and domperidone. Eight out of 13 treated patients (62%) responded to the treatment after 8 days (mean), and 5 showed a later (after $14 days) decrease in symptoms that was not attributed to the therapy. All the study patients were discharged home after spending at least 3 days free of desaturations, bradycardia, or apnea. No patients had occurrence of symptoms after the pH-MII study. All infants with BPD were discharged home in room air. Table III shows the results of multiple quantile regression analysis. In particular, orogastric tube feeding was significantly associated with acid reflux events and inversely correlated with gas reflux events. Concomitant caffeine therapy was significantly associated with a lower number of weakly acid reflux episodes. BPD had a significant positive effect on pH-only reflux events. 4
Discussion Infants with BPD did not have significantly higher GER features compared with infants without BPD as measured by esophageal pH-MII monitoring, except for higher occurrence of pH-only events and higher SSI for pH-only events.
Table III. Effect of BPD, caffeine treatment, and orogastric tube feeding on GER events: results of the quantile regression analysis Total number of reflux events BPD (yes vs no) Caffeine (yes vs no) Orogastric tube feeding (yes vs no) Total acid reflux events BPD (yes vs no) Caffeine (yes vs no) Orogastric tube feeding (yes vs no) Total weakly acid reflux events BPD (yes vs no) Caffeine (yes vs no) Orogastric tube feeding (yes vs no) Total gas reflux events BPD (yes vs no) Caffeine (yes vs no) Orogastric tube feeding (yes vs no) Total pH-only events BPD (yes vs no) Caffeine (yes vs no) Orogastric tube feeding (yes vs no)
Coefficient
95% CI
4.25 29.25 42.75
15.25; 47.11 36.82; 17.82 8.11; 70.65
1.50 8.00 11.00
12.25; 7.88 26.17; 17.99 1.49; 26.69
1.89 11.33 3.78
15.19; 4.48 24.08; 3.81 22.02; 11.83
2.15 0.77 6.85
2.18; 11.00 4.79; 5.68 9.73; 0.90
9.00 3.17 5.42
0.22; 18.07 0.41; 12.42 2.26; 23.07
Nobile et al
- 2015 pH-only events are acid reflux events lasting >5 seconds that are detected by pH-metry and not identified by MII. pH-only events may depend on immature peristalsis of the esophagus, lower LES tone, and delayed gastric empting (thus, explaining their increased frequency among preterm infants compared with older children and adults) or to other factors such as reflux esophagitis, insignificant bolus size, or slow pH drift.22,29 We speculate that: (1) infants with BPD might have impaired esophageal motility, decreased basal LES tone, and increased LES relaxations, as pointed out by others5; and (2) patients with BPD could be more sensitive than infants without BPD to the autonomic nervous system responses triggered by pH-only events that result in clinical symptoms. The latter point could be explained by the hypothesis that the esophageal afferent neural pathways might be more reactive to decreases in pH among patients with BPD because of their increased inflammatory state.30 The activation of vagal afferents could trigger airway efferent responses resulting in protective reflexes such as esophagoglottal closure and pharyngoglottal closure.31 Another explanation of our findings could rely on the fact that infants with BPD compared with infants without BPD received a significantly lower intake of fluids at the time of pH-MII because of their chronic pulmonary condition, resulting in a lower buffering capacity of milk. However, the energy intake, the type of milk and the oral lipid amount were similar in the 2 groups. It is unlikely that the feeding modality could have influenced the results because a similar proportion of patients in both groups were fed by an orogastric tube, which is a factor that could have influenced GER.9,10,29,32 The influence of the pH-MII probe itself may theoretically promote reflux; however, it is an intrinsic, unavoidable feature of the pH-MII method. The feeding duration could influence the number and characteristics of GER events.33 The feeding duration followed our general NICU policy: preterm infants received feeds every 3 hours, and each feed was administered over 20-30 minutes, with small variations depending on infants’ tolerance; it is unlikely that significant differences in feeding duration could have occurred and influenced GER features. We observed that 25% of symptoms were temporally associated with GER events. We think that the majority of symptoms in preterm infants could be explained by functional immaturity of the neuronal control of breathing and/or by autonomic system impairment because of prematurity, as pointed out by other authors.34 The temporal relationship between breathing irregularities and GER events is still controversial, as the percentage of GER-linked apnea (evaluated adopting different definitions and timing) varied from 0% to 94% in previously published studies.35-37 However, there are reports that apnea or periodic breathing in infants may lead to relaxation of LES and, thus, precede GER events.38 Although our study was not aimed at evaluating the relationship between apnea and GER, as esophageal manometry (to better evaluate LES pressure) and polysomnography (to characterize breathing irregularities) would have been needed to address this issue, we
ORIGINAL ARTICLES found that 11% of the symptoms preceded GER events. Because we could not discriminate between apnea and other symptoms, we only speculate that <10% of breathing irregularities (apnea, desaturation, and periodic breathing) may trigger LES relaxations and eventually GER episodes occurring within 120 seconds. Jadcherla et al investigated the role of GER among infants with BPD using esophageal pH-MII,14,15 and other researchers used pH-metry or clinical observations.1,9,10 Jadcherla et al also compared the occurrence of GER and symptom patterns between 18 preterm infants with BPD and 12 without BPD.14 In their work, the frequencies of GER events, acid clearance time, and SSI were all greater in infants with BPD vs those without BPD (P < .001). Higher SSI was noted with pH-only events (P < .0001 vs pH-MII events). Prolonged acid clearance times were associated with symptomatic acid reflux events in infants with BPD but not in those without BPD. However, details about respiratory management were not provided and patients did not have comparable GA at birth; these factors might explain the differences with our results. In another report, Jadcherla et al evaluated the relationship between symptoms and the most proximal extent and duration of acid reflux events in 9 premature and term infants with BPD who had combined esophageal pH-MII and manometry.15 Acid reflux events into the pharynx were associated with increased symptom occurrence and delayed clearance. Symptomatic acid reflux events were associated with longer acid clearance times that were longer the greater the extent the GER events. Overall, the esophageal acid exposure time lasted for about 15% of the day, and the distal esophagus was the most frequent site of proximal acid provocation. However, in this study, 1 patient was on nasal CPAP during the study, and 9 out of 9 were fed by a gastric tube; these factors might have influenced the GER variables.9,32 Compared with the results by Jadcherla et al, we observed that the esophageal acid exposure time was far lower (median 27 minutes per day among infants with BPD and 40.5 minutes per day among those without BPD with 210 minutes per day in the aforementioned study). There also were no differences in the SSI for proximal reflux events. An explanation of these findings could rely either on the absence of ventilatory and intensive support required by our patients or on the greater number of subjects we studied. In our study, infants with BPD compared with those without BPD had prolonged length of stay and took longer to achieve complete oral feeding. This can be explained by longer supplementary oxygen and tube feeding requirement rather than by an increased rate of GER as previously reported by others.10 We reported our experience in GER treatment with medications. The present study was not aimed at evaluating the safety and efficacy of drugs, but considering our results and the increasing concerns about the safety of antacid drugs in newborns, we believe pharmacologic treatment for GER should be initiated only after the demonstration of
Are Infants with Bronchopulmonary Dysplasia Prone to Gastroesophageal Reflux? A Prospective Observational Study with Esophageal pH-Impedance Monitoring
5
THE JOURNAL OF PEDIATRICS
www.jpeds.com
pathologic pH-MII monitoring to avoid unnecessary drug therapy, adverse events, and costs. Another finding of our study was that orogastric tube feeding was significantly associated with acid reflux events and inversely correlated with gas reflux events, as previously suggested.9,32 Finally, caffeine administration at the time of study was significantly associated with a lower number of weakly acid reflux episodes. This finding needs to be confirmed, as caffeine has been previously associated with increased gastric acid secretion and decreased LES tone.7,8 Our study has some limitations. First, there was a selection bias in testing only infants who were symptomatic. However, there are ethical difficulties precluding researchers from performing a 24-hour study in asymptomatic preterm infants to estimate GER.39 Second, although infants with BPD did not show predisposition for GER, it is important to consider that both disorders can be influenced by multiple factors (ie, prematurity per se, caffeine administration among others). Third, the study population was relatively small. To detect a difference of 8% for infants without BPD vs 15% for infants with BPD in having symptoms suggestive for GER, as found in this study (SAP $95% for weakly acid and gas reflux events), a sample of 325 infants with BPD and 325 without BPD would be needed with a power of 0.8 and alpha <.05. In conclusion, our study did not find a higher prevalence of GER features in symptomatic infants with BPD compared with symptomatic preterm infants without BPD, except for greater number of pH-only events and higher SSI for pH-only events. n We are sincerely grateful to Richard J. Grand, MD (Director Emeritus, Harvard Medical School, Boston, Massachusetts), for his valuable support and Rosaria Gesuita, PhD (Universita Politecnica delle Marche, Ancona, Italy), for her statistical contribution and insight. Submitted for publication Oct 26, 2014; last revision received Mar 25, 2015; accepted May 5, 2015.
References 1. Akinola E, Rosenkrantz TS, Pappagallo M, McKay K, Hussain N. Gastroesophageal reflux in infants <32 weeks gestational age at birth: lack of relationship to chronic lung disease. Am J Perinatol 2004;21:57-62. 2. Jobe AH. The new bronchopulmonary dysplasia. Curr Opin Pediatr 2011;23:167-72. 3. Farstad T, Bratlid D, Medbø S, Markestad T , Norwegian Extreme Prematurity Study Group. Bronchopulmonary dysplasia—prevalence, severity, and predictive factors in a national cohort of extremely premature infants. Acta Paediatr 2011;100:53-8. 4. Orenstein SR, Orenstein DM. Gastroesophageal reflux and respiratory disease in children. J Pediatr 1988;112:847-58. 5. Omari T, Barnett C, Snel A, Davidson G, Haslam R, Bakewell M, et al. Mechanism of gastroesophageal reflux in premature infants with chronic lung disease. J Pediatr Surg 1999;34:1795-8. 6. Vandenplas Y, Rudolph CD, Di Lorenzo C, Hassall E, Liptak G, Mazur L, et al. Pediatric Gastroesophageal Reflux Clinical Practice Guidelines: Joint Recommendations of the North American Society for Pediatric Gastroenterology, Hepatology, and Nutrition (NASPGHAN) and the European Society for Pediatric Gastroenterology, Hepatology, and Nutrition (ESPGHAN). J Pediatr Gastroenterol Nutr 2009;49:498-547. 6
Vol. -, No. 7. Foster LJ, Trudeau WL, Goldman AL. Bronchodilator effects on gastric acid secretion. JAMA 1979;241:2613-5. 8. Stein MR, Towner TG, Weber RW, Mansfield LE, Jacobson KW, McDonnell JT, et al. The effect of theophylline on the lower esophageal sphincter pressure. Ann Allergy 1980;45:238-41. 9. Sindel B, Maisels MJ, Ballantine TVN. Gastroesophageal reflux to the proximal esophagus in infants with bronchopulmonary dysplasia. Am J Dis Child 1989;143:1103-6. 10. Khalaf MN, Porat R, Brodsky NL, Bhandari V. Clinical correlations in infants in the neonatal intensive care unit with varying severity of gastroesophageal reflux. J Pediatr Gastroenterol Nutr 2001;32:45-9. 11. Giuffre RM, Rubin S, Mitchell I. Antireflux surgery in infants with bronchopulmonary dysplasia. Am J Dis Child 1987;141:648-51. 12. Hoyoux C, Forget P, Lambrechts L, Geubelle F. Chronic bronchopulmonary disease and gastroesophageal reflux in children. Pediatr Pulmonol 1985;3:149-53. 13. St Cyr JA, Ferrara TB, Thompson T, Johnson D, Foker JE. Treatment of pulmonary manifestations of gastroesophageal reflux in children two years of age or less. Am J Surg 1989;157:400-4. 14. Jadcherla SR, Peng J, Chan CY, Moore R, Wei L, Fernandez S, et al. Significance of gastroesophageal refluxate in relation to physical, chemical and spatio-temporal characteristics in symptomatic ICU neonates. Pediatr Res 2011;70:192-8. 15. Jadcherla S, Gupta A, Fernandez S, Nelin LD, Castile R, Gest AL, et al. Spatiotemporal characteristics of acid refluxate and relationship to symptoms in premature and term infants with chronic lung disease. Am J Gastroenterol 2008;103:720-8. 16. Terrin G, Passariello A, De Curtis M, Manguso F, Salvia G, Lega L, et al. Ranitidine is associated with infections, necrotizing enterocolitis, and fatal outcome in newborns. Pediatrics 2012;129:e40-5. 17. Wenzl TG, Silny J, Schenke S, Peschgens T, Heimann G, Skopnik H. Gastroesophageal reflux and respiratory phenomena in infants: status of the intraluminal impedance technique. J Pediatr Gastroenterol Nutr 1999;28:423-8. 18. Di Fiore JM, Arko MK, Miller MJ, Krauss A, Betkerur A, Zadell A, et al. Cardiorespiratory events in preterm infants referred for apnea monitoring studies. Pediatrics 2001;108:1304-8. 19. Walsh MC, Yao Q, Gettner P, Hale E, Collins M, Hensman A, et al. Impact of a physiologic definition on bronchopulmonary dysplasia rates. Pediatrics 2004;114:1305-11. 20. Condino AA, Sondheimer J, Pan Z, Gralla J, Perry D, O’Connor JA. Evaluation of infantile acid and nonacid gastroesophageal reflux using combined pH monitoring and impedance measurement. J Pediatr Gastroenterol Nutr 2006;42:16-21. 21. Wenzl TG, Benninga MA, Loots CM, Salvatore S, Vandenplas Y, Diaz JA, et al. Indications, methodology, and interpretation of combined esophageal impedance-pH monitoring in children: ESPGHAN EURO-PIG standard protocol. J Pediatr Gastroenterol Nutr 2012;55:230-4. 22. Shin MS, Shim JO, Moon JS, Kim HS, Ko JS, Choi JH, et al. ImpedancepH monitoring and conventional pH monitoring are complementary methods to detect association between gastroesophageal reflux and apnea-related symptoms in preterm infants and neonates. J Matern Fetal Neonatal Med 2012;25:2406-10. 23. Shay S, Tutuian R, Sifrim D, Vela M, Wise J, Balaji N, et al. Twenty-four hour ambulatory simultaneous impedance and pH monitoring: a multicenter report of normal values from 60 healthy volunteers. Am J Gastroenterol 2004;99:1037-43. 24. Sifrim D, Mittal R, Fass R, Smout A, Castell D, Tack J, et al. Acidity and volume of the refluxate in the genesis of gastro-oesophageal reflux disease symptoms. Aliment Pharmacol Ther 2007;25:1003-17. 25. Pilic D, Fr€ ohlich T, N€ oh F, Pappas A, Schmidt-Choudhury A, K€ ohler H, et al. Detection of gastroesophageal reflux in children using combined multichannel intraluminal impedance and pH measurement: data from the German Pediatric Impedance Group. J Pediatr 2011;158:650-4. 26. Breumelhof R, Smout AJ. The symptom sensitivity index: a valuable additional parameter in 24-hour esophageal pH recording. Am J Gastroenterol 1991;86:160-4.
Nobile et al
- 2015 27. Weusten BL, Roelofs JM, Akkermans LM, Van Berge-Henegouwen GP, Smout AJ. The symptom-association probability: an improved method for symptom analysis of 24-hour esophageal pH data. Gastroenterology 1994;107:1741-5. 28. Koenker R. Quantile regression. New York: Cambridge University Press; 2005. 29. Di Fiore JM, Arko M, Churbock K, Hibbs AM, Martin RJ. Technical Limitations in detection of gastroesophageal reflux (GER) in neonates. J Pediatr Gastroenterol Nutr 2009;49:177-82. 30. Bhandari V. Postnatal inflammation in the pathogenesis of bronchopulmonary dysplasia. Birth Defects Res A Clin Mol Teratol 2014;100:189-201. 31. Jadcherla SR, Gupta A, Wang M, Coley BD, Fernandez S, Shaker R. Definition and implications of novel pharyngo-glottal reflex in human infants using concurrent manometry ultrasonography. Am J Gastroenterol 2009;104:2572-82. 32. Solana Garcıa MJ, L opez-Herce Cid J, Sanchez Sanchez C. Gastroesophageal reflux in critically ill children: a review. ISRN Gastroenterol 2013; 2013:824320. 33. Jadcherla SR, Chan CY, Moore R, Malkar M, Timan CJ, Valentine CJ. Impact of feeding strategies on the frequency and clearance of acid
ORIGINAL ARTICLES
34.
35. 36.
37.
38. 39.
and nonacid gastroesophageal reflux events in dysphagic neonates. JPEN J Parenter Enteral Nutr 2012;36:449-55. Eichenwald EC, Aina A, Stark AR. Apnea frequently persists beyond term gestation in infants delivered at 24 to 28 weeks. Pediatrics 1997; 100:354-9. Arad-Cohen N, Cohen A, Tirosh E. The relationship between gastroesophageal reflux and apnea in infants. J Pediatr 2000;137:321-6. Mousa H, Woodley FW, Metheney M, Hayes J. Testing the association between gastroesophageal reflux and apnea in infants. J Pediatr Gastroenterol Nutr 2005;41:169-77. Peter CS, Sprodowski N, Bohnhorst B, Silny J, Poets CF. Gastroesophageal reflux and apnea of prematurity. No temporal relationship. Pediatrics 2002;109:8-11. Omari TI. Apnea-associated reduction in lower esophageal sphincter tone in premature infants. J Pediatr 2009;154:374-8. Vandenplas Y, Belli D, Boige N, Bouquet J, Cadranel S, Cezard JP, et al. A standardized protocol for the methodology of esophageal pH monitoring and interpretation of the data for the diagnosis of gastroesophageal reflux. J Pediatr Gastroenterol Nutr 1992;14: 467-71.
Are Infants with Bronchopulmonary Dysplasia Prone to Gastroesophageal Reflux? A Prospective Observational Study with Esophageal pH-Impedance Monitoring
7
THE JOURNAL OF PEDIATRICS
www.jpeds.com
Vol. -, No. -
Figure. A, 24 hour pH-MII recording. B, Detail of a pH-only event that was followed by a symptom.
7.e1
Nobile et al