online 11 Sep 2008;
2009;94;F188-F192; originally publishedArch. Dis. Child. Fetal Neonatal Ed.
L Corvaglia, D Zama, S Gualdi, M Ferlini, A Aceti and G Faldella
of apnoeas in very preterm infants
Gastro-oesophageal reflux increases the number
Updated information and services can be found at:
This article cites 23 articles, 6 of which can be accessed free at:
You can respond to this article at:
the top right corner of the article
Receive free email alerts when new articles cite this article - sign up in the box at
To order reprints of this article go to:
go to: Archives of Disease in Childhood - Fetal and Neonatal EditionTo subscribe to
on 27 April 2009
Gastro-oesophageal reflux increases the number of
apnoeas in very preterm infants
L Corvaglia, D Zama, S Gualdi, M Ferlini, A Aceti, G Faldella
Neonatology and Neonatal
Intensive Care Unit, University of
Bologna, Bologna, Italy
Luigi Corvaglia, Neonatologia e
Terapia Intensiva Neonatale, Via
Massarenti, 11. 40138 Bologna,
Accepted 11 August 2008
Published Online First
11 September 2008
Objective: To document the existence of a relationship
between apnoea of prematurity (AOP) and gastro-
oesophageal reflux (GER) in preterm infants.
Setting: Neonatal intensive care unit.
Patients: Twenty-six preterm infants (gestational age
(32 weeks) with recurrent apnoeas.
Intervention: Simultaneous and synchronised recording
of polysomnography and pH-impedance monitoring (pH-
MII). Polysomnography detects and characterises
apnoeas, by recording of breathing movement, nasal
airflow, electrocardiogram and pulse oximeter saturation.
pH-MII is the state-of-the-art methodology for GER
detection in preterm newborns.
Main outcome measures: Relationship between AOP
and GER, which were considered temporally related if
both started within 30 seconds of each other.
Results: One hundred and fifty-four apnoeas out of 1136
were temporally related to GER. The frequency of apnoea
during the 1-minute time around the onset of GER was
significantly higher than the frequency detected in the
GER-free period (p=0.03). Furthermore, the frequency of
apnoea in the 30 seconds after GER (GER-triggered
apnoeas) was greater than that detected in the
30 seconds before (p=0.01). A great inter-individual
variability was documented in the proportion of GER-
triggered apnoeas. A strong correlation between total
number of apnoeas and the difference between apnoeas
detected 30 seconds after and before GER was found
Conclusions: Our data show that a variable rate of
apnoeas can be triggered by GER in very preterm infants.
Further studies are needed to recognise clinical features
that identify those patients who are more susceptible to
Apnoea of prematurity (AOP) is a perplexing
disorder of respiratory control that is very common
in preterm infants and requires therapeutic inter-
vention to avoid potential morbidity.1The diag-
nosis of AOP is made by exclusion, when other
possible infectious, cardiologic, physiological and
metabolical causes of apnoea have been ruled out.2
The existence of a relationship between gastro-
oesophageal reflux (GER) and AOP in preterm
infants is still controversial.3This hypothesis is
suggested by the clinical observation that both
GER and apnoeas frequently occur during post-
prandial periods. In clinical practice, many preterm
infants receive pharmacological treatment for GER,
with the expectation that treatment would reduce
the frequency and severity of episodes of apnoea.4 5
Furthermore, the hypothesis of a linkage is
supported by experimental studies performed on
preterm animal newborns, whose results have also
been confirmed in human preterm infants.6–9
At present, clinical studies aiming to describe a
linkage between GER and AOP have shown
contrasting results. While Wenzl and Menon
showed an association between GER and respira-
tory phenomena in infancy,10 11other authors
excluded a temporal relationship between GER
and apnoeas both in infants12and in preterm
We aimed to evaluate the relationship between
GER and apnoeas in preterm newborns, by the
simultaneous and synchronised use of two differ-
ent techniques: combined multichannel intralum-
inal impedance and pH monitoring (pH-MII), and
detailed detection of both acid and non-acid GER
episodes, while the second detects apnoeas and
characterises them as central (CA), obstructive
(OA) and mixed (MA).
Preterm newborns with gestational age (GA)
(32 weeks were eligible if they had recurrent
apnoeas (at least two episodes of apnoeas, docu-
mented by nursing charts, over a 2 h period in the
2 days before the study13), and were on full enteral
feeding. Newborns suffering from sepsis, necrotis-
ing enterocolitis, intra-ventricular haemorrhage or
major congenital abnormalities were ruled out.
Each infant underwent a 6 h (two post-prandial
periods lasting 3 h each) recording of simultaneous
PSG and pH-MII; the clocks of the two instru-
ments were synchronised at minute–second by one
What is already known on this topic
Apnoea of prematurity (AOP) and gastro-oesopha-
geal reflux (GER) are both common in preterm
infants. At present, studies aiming to establish a
relationship between AOP and GER have yielded
What this study adds
Gastro-oesophageal reflux (GER) can act as a
trigger mechanism for apnoeas in preterm infants.
The proportion of apnoeas that are triggered by GER
is extremely variable: the higher the number of
apnoeas, the higher the number of GER-triggered
F188Arch Dis Child Fetal Neonatal Ed 2009;94:F188–F192. doi:10.1136/adc.2008.143198
on 27 April 2009
of the investigators before starting each recording. The
recording of the two layouts was started simultaneously by
one of the investigators. Written informed parental consent was
obtained before each examination.
pH-MII monitoring allows a detailed detection of both acid and
non-acid GER. MII recognises a GER as a change in electrical
impedance which occurs during bolus passage inside the
oesophagus. Variations in impedance are sequentially measured
by closely arranged electrodes and the direction of bolus is
determined by the temporal sequence of impedance changes
recorded by the different electrodes (bolus moving in cranial–
caudal direction is recognised as a swallow, whereas bolus
moving in the opposite direction as a reflux14).
The single-use, flexible MII-pH probe (ComfortecHMII-pH
Sandhill Scientific Highlands Ranch, Colorado, USA) contains
seven impedance rings, which constitute six 1.5 cm dipolar
impedance channels, and one antimony electrode for pH
detection, located in the middle of the distal impedance dipole.
The catheter is inserted through a nostril, without sedation, and
is placed under fluoroscopy. The tip is fixed about 1 cm above
the lower oesophageal sphincter (LES). Before removal, the
position of the catheter is compared with the initial position to
exclude a possible displacement. Data are acquired on a portable
Sleuth Sandhill Scientific system, and then stored in a personal
computer; they are analysed by specific software (BioVIEW
Analysis, Sandhill Scientific, version 5.0.9) and confirmed by
direct visual evaluation of the layout.
To be detected automatically as a GER by MII software, a
reflux episode must reach at least three consecutive rings,
causing a drop in impedance .50% from baseline in each dipole.
For this reason, MII automatic scan is ‘‘blind’’ to all GER
episodes shorter than 4 cm. This is particularly relevant in
preterm newborns, due to the disproportion between their
oesophageal length (approximately 8 cm) and the distance
between the rings inside the MII catheter.
On the other hand, pH monitoring recognises all acid GER
episodes reaching the pH electrode (about 2 cm above LES) as a
drop in oesophageal pH below 4. In conclusion, MII automatic
scan is able to detect all GER episodes reaching the second
impedance dipole, which are defined as acid and non-acid
MII-GER (pH lower or higher than 4, respectively). pH
monitoring also records acid GER episodes limited to distal
oesophagus, called pH GER.15–17The evaluation of non-acid
GER episodes which do not reach the second impedance dipole
(short non-acid MII-GER) must be performed by a visual
reading of the layout. In the present study, a short non-acid
GER has been defined as a drop in impedance .50% from
baseline in the distal dipole, in the absence of any impedance
change in the other dipoles.
PSG is the most adequate electrophysiological test to identify
apnoea in preterm and term newborns.18Layouts, recorded by a
computerised PSG (Micromed System Plus), are acquired using
different devices: a flow transducer to measure oral/nasal
airflow; a band placed around the abdomen, which constitutes
a respiratory inductance plethysmography, to evaluate abdom-
inal movements; a pulse oximeter sensor, placed on the infant’s
hand or foot, to measure the amount of haemoglobin which is
saturated with oxygen; and two electrodes placed on the thorax
under the clavicles, to register cardiac electrical activity. A
portable camera documents any possible artefact (movements
during crying, electrode displacement, etc).
Each infant was observed by an investigator during the
monitoring to record any change in infant behaviour and to
avoid any technical artefact of the layout. No sedation or sleep
deprivation was used.
Each layout was analysed visually by one of the investigators.
The Sleep Analysis software provided by the manufacturer was
not relied upon because it has not been validated in preterm
We considered as apnoeas each oral/nasal flow cessation
lasting at least 5 seconds. Apnoeas were then divided into
central (CA: absence of oral/nasal flow without associated
respiratory effort documented by the absence of movements of
abdominal walls), obstructive (OA: respiratory abdominal
movements with an inadequate oral/nasal flow) and mixed
apnoeas (MA: components of both CA and OA). All the apnoeas
lasting >20 seconds, or lasting >5 seconds and followed by
desaturations (SatO2,85%) and/or bradycardia (heart frequency
,100 bpm), were defined as pathological.19Apnoeas included in
periodic breathing,20or apnoeas, desaturations and bradycardia
occurring during meals were ruled out from the analysis.
Correlation between GER and apnoea
The temporal association between GER and apnoea was
analysed by the frequency of apnoea in the 30 seconds
preceding and following the onset of GER episodes. Apnoeas
detected within 30 seconds after GER were defined as GER-
All statistical analyses were performed by SPSS 16.0 (Statistical
Package for the Social Sciences SPSS Inc., version 16.0) for
Windows. The association between GER and apnoeas was
tested by Wilcoxon-signed rank test. The correlation between
the total number of apnoeas and GER-related apnoeas was
checked by Spearman test. A value of p,0.05 was considered
Twenty-six preterm infants (eight male, 18 female) with a
median GA of 30 weeks (range 25–32 weeks) and a mean body
weight (BW) of 1247 g were studied at a median GA of
34 weeks and a mean BW of 1704 g. Eight infants still needed
supplemental oxygen, seven were receiving caffeine and one
Doxapram. All were on full enteral feeding: 13 were fed by a
feeding tube, removed after a meal, while the remaining 12 were
bottle fed. None of them had malformation or major gastro-
intestinal problems, or were taking drugs influencing gastro-
intestinal motility or gastric acidity. All the patients tolerated
the test well and their clinical conditions remained stable. For
technical reasons, one patient had to be excluded from the
The investigator who analysed pH-MII layouts was blind to
the results of the PSG, and vice versa.
During 151 h of registration (mean 6 h and 2 minutes for each
study), 1065 GER episodes (mean 42/newborn, range 15–93)
were identified; 382 (35.9%) of them were detected only by pH
monitoring, and therefore classified as pH GER. The remaining
683 (64.1%) were detected by MII, and further divided into acid
Arch Dis Child Fetal Neonatal Ed 2009;94:F188–F192. doi:10.1136/adc.2008.143198F189
on 27 April 2009
MII-GER (n=120, 11.3%), non-acid MII-GER (n=488, 45.8%)
and short non-acid MII-GER (n=75, 7%).
One thousand one hundred and thirty-six apnoeas (mean 45/
newborn, range 1–169), 342 desaturations (mean 13.7/newborn,
range 0–104) and 37 bradycardias (mean 1.5/newborn, range
0–14) were detected. Five hundred and two (44.2%) apnoeas
were central, 194 (17.1%) obstructive and 440 (38.7%) mixed.
One hundred and fifty-six (13.7%) apnoeas were pathological,
and 980 (86.3%) were non-pathological.
One hundred and fifty-four apnoeas out of 1136 were detected
within 30 seconds before and/or after the onset of a GER
episode. For each patient, the frequency of apnoea during the
1-minute time around the onset of a GER episode was
calculated by dividing the number of apnoeas detected in that
time interval by the cumulative time of those 1-minute time
As shown in fig 1, the frequency of apnoea during the
1-minute time around the onset of a GER episode (mean 0.17/
min; range 0–0.84/min) was significantly different (p=0.03)
from the one detected during the GER-free period (0.12/min (0–
0.43/min)), which is defined as the time difference between the
total recording time and the time around all GER episodes. The
frequency of apnoea before the onset of GER (0.10/min (0–0.69/
min)) did not differ from the frequency detected during the
GER-free period (p=0.17). The frequency of apnoea occurring
after the onset of GER (0.25/min (0–1/min)) was higher than
the frequency detected in the GER-free period (p=0.02), and
also than the one detected before the onset of GER (p=0.01).
A great inter-individual variability was documented in the
ratio of GER-triggered apnoeas to the total number of apnoeas
(median 10.34%, range 0–20% (fig 2)), as well as in the ratio of
GER-triggered apnoeas to the total number of GERs (median
5.77%, range 0–50%).
The increase of apnoeas after GER, defined as the difference
between the number of apnoeas detected in the 30 seconds after
GER and the number detected 30 seconds before GER, was
calculated (median 2, range 22/8). A significant correlation was
found between this difference and the total number of apnoeas
(p=0.425, p=0.034 (fig 3)); on the other hand, no correlation
was documented between total number of GERs and GER-
triggered apnoeas (p=0.019, p=0.930).
No correlation was found between GER-triggered apnoeas
and gestational age, postmenstrual age, birth weight, oxygen
requirement, feeding characteristics and treatment with caf-
Our study was drawn to investigate the controversial correla-
tion between AOP and GER by means of the state-of-the-art
techniques for the detection of these two events.
We found that, in preterm infants with recurrent apnoeas,
these events occur more frequently soon after GER than in the
period immediately before GER or the GER-free period. This
suggests that the relationship between GER and AOP is not
simply by chance; if both events would be related only to
immaturity, there would be no difference in the frequency of
apnoeas detected before and after GER.
The ratio of GER-triggered apnoeas to the total number of
apnoeas is widely variable: while in some newborns the
percentage of GER-triggered apnoeas is very small, in others
this percentage is relevant. Furthermore, the wide variability in
the ratio of GER-triggered apnoeas to the total number of GERs
documents that in some newborns a GER episode more likely
evokes an apnoea.
Patients who had the most significant increase of apnoeas
after GER also had the highest number of total apnoeas, that is,
in the most unstable infants, who often experience apnoeas,
GER episodes can easily act as a trigger event for apnoeas. On
the other hand, the lack of correlation between GER-triggered
apnoeas and total number of GERs highlights that the severity
of GER is not predictive of the risk of reflex apnoeas. In our
study population, no clinical feature was found to be related to
an increased risk of GER-triggered apnoeas.
The strength of our study is given by the methodological
accuracy in the detection of both GER and apnoeas: PSG allows
a precise evaluation of each kind of apnoea (CA, OA and MA),
and combined pH-MII offers a detailed description of both acid
and non-acid GER episodes. This is particularly relevant for
preterm newborns, whose gastric acidity is often buffered by
frequent milk meals. A limit of combined pH-MII in our study
population may be related to the software’s inability to detect
short non-acid GER automatically: an improvement in the
characteristics of the software or the catheter is advisable to
further improve the methodology. In our study, we have
overcome this limit by the direct and visual analysis of all the
layouts. Finally, the synchronisation of the two techniques at
(number per minute) in the gastro-
oesophageal reflux (GER)-free period, in
the 1-minute time around GER,
30 seconds before GER and 30 seconds
after GER. Values are expressed as mean
Frequency of apnoeas
F190Arch Dis Child Fetal Neonatal Ed 2009;94:F188–F192. doi:10.1136/adc.2008.143198
on 27 April 2009
minute–second has granted a precise evaluation of the temporal
relationship between the events.
Previously, experimental studies performed on animal models
have shown that the infusion of a liquid solution into the larynx
or oesophagus evokes the so-called chemolaryngeal reflex, which
can cause apnoeas and swallows.6–8Jadcherla described a reflex
evoked by water stimulation in the mid-oesophagus and
characterised by the occurrence of central apnoeas, swallows
and alteration of oesophageal motility.9During fetal life and
after birth, this reflex is supposed to be protective towards the
inhalation of gastric content into the airways, and seems to
develop into a cough reflex.6In fact, recurrent apnoeas and
swallows are frequently observed in preterm newborns, while in
term newborns apnoeas are generally shorter and less frequent.
Our data are in contrast with recent findings by Peter13and Di
Fiore,21who excluded a temporal relationship between GER and
apnoea in preterm newborns. In our opinion, the explanation of
this discrepancy is methodological, as we did and they did not
use combined pH-MII, which is considered the state-of-the-art
methodology to detect GER.
The absence of a pH sensor in the study by Peter did not
allow the detection of acid GER episodes limited to the distal
oesophagus. It was previously demonstrated that these episodes
can evoke an apnoea by a reflex mechanism,7and in our study
they were found to represent a great proportion (42.9%) of the
total amount of GER. For both these reasons, their detection
cannot be left out in preterm newborns. Furthermore, these
authors used a trans-LES catheter which can increase the
number of GERs, as the same authors previously demon-
Di Fiore et al evaluated the association of GER and AOP by
means of pH monitoring alone, which is not considered as
accurate as combined pH-MII for GER detection in preterm
newborns.23 24pH monitoring alone cannot detect non-acid GER
episodes, whose number is high in preterm newborns, as shown
also in our study. Furthermore, these authors recorded apnoeas
without flow sensor, and that leads to an obvious risk of
underestimating OAs and MAs.
Only few studies aiming to correlate GER and apnoeas were
performed by simultaneous PSG and combined pH-MII;
although performed in term infants, they have shown a
temporal association between apnoea and GER.10
A possible limit of this study can be related to the arbitrary
choice of the time interval used to establish a temporal
correlation between GER and AOP. However, even though no
pathophysiological data support the use of this time interval,
the evaluation of 30 seconds before and after the onset of each
episode is the one mostly reported in clinical studies.
Another limit is related to our relatively small study
population, which did not allow us to identify specific
30 seconds after gastro-oesophageal
reflux (GER) to the total number of
apnoeas for each patient. Values are
expressed as percentages.
Ratio of apnoeas detected
number of apnoeas and the difference
between apnoeas detected 30 seconds
after gastro-oesophageal reflux (GER) and
30 seconds before GER.
Correlation between total
Arch Dis Child Fetal Neonatal Ed 2009;94:F188–F192. doi:10.1136/adc.2008.143198F191
on 27 April 2009
characteristics and/or comorbidities which can lead to an Download full-text
increased risk of GER-triggered apnoeas.
In conclusion, this is the first study using the state-of-the-art
methodology for the simultaneous and synchronised recording
of GER and apnoeas in preterm infants. Our data show that a
variable rate of apnoeas can be triggered by GER in very preterm
infants. Further studies are needed to recognise clinical features
that identify those patients who are more susceptible to GER-
Competing interests: None.
Patient consent: Parental consent obtained.
Miller MJ, Martin RJ. Pathophysiology of apnea of prematurity. In: Polin RA, Fox
WW, Abman SH, eds. Fetal and neonatal physiology. 3rd edn. Philadelphia, USA:
Elsevier Inc, 2004: 905–17.
Bhatia J. Current options in the management of apnea of prematurity. Clin Pediatr
Slochum C, Hibbs AM, Martin RJ, et al. Infant apnea and gastroesophageal reflux: a
critical review and framework for further investigation. Curr Gastroenterol Rep
Bancalari E. Apnea is not prolonged by acid gastroesophageal reflux in preterm
infants. Pediatrics 2005;116:1217–18.
Dhillon AS, Ewer AK. Diagnosis and management of gastroesophageal reflux in
preterm infants in neonatal intensive care units. Acta Pediatr 2004;93:88–93.
Thach BT. Maturation and transformation of reflexes that protect the laryngeal
airway from liquid aspiration from fetal to adult life. Am J Med 2001;111:69s–77s.
Perkett EA, Vaughan RL. Evidence for a laryngeal chemoreflex in some human
preterm infants. Acta Paediatr Scand 1982;71:969–72.
Pickens DL, Schefft G, Thach BT. Prolonged apnea associated with upper airway
protective reflexes in apnea of prematurity. Am Rev Respir Dis 1988;137:113–18.
Jadcherla SR. Upstream effect of esophageal distension: effect on airway. Curr
Gastr Rep 2006;8:189–93.
Wenzl TG, Schenke S, Peschgens T, et al. Association of apnea and non acid
gastroesophageal reflux in infants: investigations with the intraluminal impedance
technique. Pediatr Pulmonol 2001;31:144–9.
Menon PA, Schefft GL, Thach BT. Apnea associated with regurgitation in infants.
J Pediatr 1985;106:625–9.
Mousa H, Woodley FW, Metheney M, et al. Testing the association between
gastroesophageal reflux and apnea in infants. J Pediatr Gastroenterol Nutr
Peter CS, Sprodowski N, Bohnhorst B, et al. Gastroesophageal reflux and apnea of
prematurity: no temporal relationship. Pediatrics 2002;109:8–11.
Wenzl TG. Investigating esophageal reflux with the intraluminal impedance
technique. J Pediatr Gastroenterol Nutr 2002;34:261–8.
Sifrim D, Castell D, Dent J, et al. Gastro-esophageal reflux monitoring: review and
consensus report on detection and definitions of acid, non acid and gas reflux. Gut
Corvaglia L, Ferlini M, Rotatori R, et al. Starch thickening of human milk is ineffective
in reducing the gastroesophageal reflux in preterm infants: a crossover study using
intraluminal impedance. J Pediatr 2006;148:265–8.
Corvaglia L, Rotatori R, Ferlini M, et al. The effect of body positioning on
gastroesophageal reflux in premature infants: evaluation by combined impedance and
pH monitoring. J Pediatr 2007;151:591–6.
American Thoracic Society. Standards and indications for cardiopulmonary sleep
studies in children. Am J Respir Crit Care Med 1996;153:866–78.
Di Fiore JM, Arko MK, Miller MJ, et al. Cardiorespiratory events in preterm infants
referred for apnea monitoring studies. Pediatrics 2001;108:1304–8.
Kelly DH, Stellwagen LM, Kaitz E, et al. Apnea and periodic breathing in normal full
term infants during the first twelve months. Pediatr Pulmonol 1985;1:215–19.
Di Fiore JM, Arko MK, Whitehouse M, et al. Apnea is not prolonged by acid
gastroesophageal reflux in preterm infants. Pediatrics 2005;116:1059–63.
Peter CS, Wiechers C, Bohnorst B, et al. Influence of nasogastric tube on
gastroesophageal reflux in preterm infants: A multiple intraluminal impedance study.
J Pediatr 2002;141:277–9.
Grant L, Cochran D, Ewer AK. Can pH monitoring reliably detect gastro-oesophageal
reflux in preterm infants? Arch Dis Child Fetal Neonatal Ed 2001;85:F155–8.
Wenzl TG, Skopnik H. Intraluminal impedance: an ideal technique for evaluation
of pediatric gastroesophageal reflux disease. Curr Gastroenterol Rep
BMJ Masterclasses are educational meetings designed specifically to meet the learning needs of
doctors. They help doctors keep up to date with the latest evidence and recent guidelines in major
clinical areas, enabling them to use the latest evidence to make better decisions. The latest evidence,
recent guidelines and best practice are delivered in an interactive and informative manner by leading
experts. The speakers are specifically chosen as highly-skilled communicators who can authoritatively
enthuse the audience and interpret the latest research and guidelines into practical tips for busy
doctors. BMJ Masterclasses have proved a huge hit with clinicians, with many saying they have
influenced their clinical practice.
F192Arch Dis Child Fetal Neonatal Ed 2009;94:F188–F192. doi:10.1136/adc.2008.143198
on 27 April 2009