 |
|
 |
|
 |
 |
|
 |
 |
 |
 |
 |
 |
 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
From the Division of Neonatology, Department of Pediatrics, Rainbow Babies and Children’s Hospital, Case Western Reserve University, Cleveland, Ohio
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Objective. To characterize apnea, bradycardia, and desaturation events in infants referred for an overnight apnea monitoring study and compare them with corresponding events in control infants of similar age and weight with no bedside monitor alarms.
Methods. Twelve-hour bedside apnea monitoring studies were performed on 68 preterm infants before hospital discharge. This population included 35 infants who were referred by their attending physicians because of persistent bedside monitor alarms (referral group) and 33 infants who had no documented cardiorespiratory events for at least 2 days before the study (control group). Each study monitored respiration via respiratory inductance plethysmography, oxygen saturation (Sao2), and heart rate. Events were defined as meeting 1 of the following criteria: apnea 
20 seconds, bradycardia 
80 beats per minute, or Sao2 80%.
Results. The incidence of apnea 20 seconds was low, with no significant difference between infant groups. Referral infants exhibited a higher occurrence of desaturation episodes (20 ± 6 vs 6 ± 3 episodes/12-hour study) and a higher occurrence of bradycardia episodes (4.3 ± 0.8 vs 1.1 ± 0.3 episodes/12-hour study) than controls. These episodes of desaturation and bradycardia were always preceded by a respiratory pause, which was shorter in the referral infants (10.0 ± 0.4 seconds vs 12.0 ± 1.0 seconds). Baseline Sao2 was lower in referrals than controls (95 ± 1% vs 98 ± 1%), and the incidence of periodic breathing was significantly higher.
Conclusions. Infants referred for apnea monitoring studies because of persistent bedside monitor alarms have very infrequent prolonged apnea but a higher frequency of desaturation and bradycardia in response to short respiratory pauses than infants without persistent bedside monitor alarms. Referral infants also exhibit a lower baseline Sao2. These abnormalities in oxygenation and cardiorespiratory control may be markers for subtle residual lung disease or functional central nervous system abnormalities.
Key Words: apnea • bradycardia • oxygen saturation
Abbreviations: RIP, respiratory inductance plethysmography • Sao2, oxygen saturation
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
During hospitalization, premature infants who have persistent bedside monitor alarms often are referred for an apnea monitoring study. At our institution this study includes continuous measurement of respiration via respiratory inductance plethysmography (RIP), oxygen saturation (Sao2), and heart rate performed overnight at the infant’s bedside. The results are designed to assist the clinician with decisions about both hospital discharge and the need for home monitoring. Therefore, the purpose of this study was to characterize and compare apnea, bradycardia, and desaturation events recorded during an overnight study in infants referred for cardiorespiratory events (referral group) with corresponding events in infants of similar age and weight with no bedside monitor alarms (control group). We hypothesized that this monitoring protocol would identify infants with subtle residual abnormalities of cardiorespiratory control.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
20 seconds, bradycardia 80 beats per minute, and saturation 80% when pulse oximetry was used. The most common reason for referral was recurrent bradycardia episodes, with 94% of infants having such events. Additionally, 36% of the infants had desaturation episodes, although the application and duration of clinical pulse oximetry monitoring were inconsistent, with 90% of infants having continuous bedside Sao2 monitoring. Infants with bedside alarms occurring only during feeding were excluded from the study. Comparable studies were performed in 33 control infants who had no clinically documented cardiorespiratory events for at least 2 days before the study and were preparing for hospital discharge. Infants were enrolled at Rainbow Babies & Children’s Hospital over a period of 2 years (1998 to 2000). The study was approved by the institutional review board, and written informed parental consent was obtained before enrollment in the study. Inclusion criteria for both groups were birth weight <1800 g, postconceptional age between 33 and 36 weeks at the time of study, and no theophylline for at least 72 hours before the study. Exclusion criteria included the presence of congenital anomalies, major cardiac disorder, or need for supplemental O2, bronchodilator, or diuretic therapy. Infants with known antenatal exposure to illicit drugs or history of maternal alcohol abuse were also excluded. Demographic data are presented in Table 1.
|
After each study was completed, the data files were downloaded to a desktop computer and events were scored visually using RespiEvents Software (Sensormedics, Yorba Linda, CA). Based on widely accepted and previously published threshold settings,3,6,9,10,11 events were defined as meeting at least 1 of the following criteria: apnea 20 seconds, bradycardia 80 beats per minute, or Sao2 80%.
The total duration of the respiratory pause accompanying each event was recorded using the tidal volume wave form. In addition, the duration of the central and obstructed components of each event were recorded, with the obstructed component defined as the occurrence of asynchronous rib cage and abdominal movement during periods of absent (zero) tidal volume.
Baseline and lowest values during each event were recorded for both Sao2 and heart rate. Baseline values were based on a 30-second average of artifact-free wave forms immediately before the event. Artifact for the Sao2 wave form was identified via pulse wave form. An artifact-free signal was defined as having a clear pulse for every corresponding R wave on the electrocardiogram tracing in the absence of low-frequency oscillations indicating body movement. The following timing parameters were also measured: time delay from beginning of the respiratory pause to onset of both the heart rate deceleration and desaturation and time delay from the beginning of the respiratory pause to the lowest heart rate and saturation.
Because periodic breathing often is associated with short episodes of bradycardia and desaturation, the number of periodic breathing cycles was documented for each 12-hour study. Periodic breathing cycles were defined as repeated ventilatory cycles with a clear respiratory pause of at least 2 seconds between cycles.12 This distinguished periodic breathing from isolated apnea, allowing each event to be classified as associated with periodic breathing or isolated apnea.
Comparisons between groups were made via unpaired t test, Mann-Whitney U test (where indicated) for unpaired data not showing a Gaussian distribution, and 
2 where appropriate. A P value of <.05 was needed for significance. Data are presented as mean ± standard error of the mean.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
20 seconds was low and occurred in 8 of 35 referral infants and 4 of 33 control infants, with no significant difference between infant groups (0.7 ± 0.4 vs 0.3 ± 0.2 apneas/12-hour study, referrals vs controls, respectively; NS Mann–Whitney; Table 2). Although it was expected that apnea 20 seconds would be accompanied by desaturation or bradycardia, this was not always the case. Five apneic events (range 23 to 65 seconds) were recorded in 1 referral infant who exhibited neither desaturation nor bradycardia as defined by this study, although an additional 65 events in this same infant did result in bradycardia 80 beats per minute or saturation 80%. Average baseline heart rate for this infant was 157 beats per minute, slightly less than average for the referral group, and baseline Sao2 for this infant was 99%, which fell in the upper quartile for this infant group.
|
80% occurred in 33 of 35 referral infants but only 19 of 33 control infants (19.7 ± 5.6 vs 5.5 ± 2.6 events/12-hour study, respectively; P = .003 Mann–Whitney; Table 2). Bradycardia 80 beats per minute occurred in 26 of 35 referral infants and 15 of 33 control infants, also with a higher mean incidence in the referral infants (4.3 ± 0.9 vs 1.1 ± 0.3 events/12-hour study, respectively; P < .003 Mann–Whitney; Table 2).
Although prolonged apnea 20 seconds occurred infrequently, all desaturation and bradycardia episodes were associated with a respiratory pause. The duration of this respiratory pause was shorter for the referral infants than the control infants (10.0 ± 0.4 seconds vs 12.0 ± 1.0 seconds, respectively; P < .05). For the referral and control infants, respectively, 64% versus 51% of these respiratory pauses were central in origin, and 10% versus 8% of these respiratory pauses were purely obstructive in origin. The remaining respiratory pauses were mixed in origin and consisted of an initial central component (7.0 ± 0.9 vs 7.2 ± .7 seconds duration, referral vs control) followed by obstructive efforts (4.5 ± 0.4 vs 5.0 ± 0.6 seconds, referral vs control) before resolution of the respiratory pause. Desaturation episodes were evenly distributed between central, obstructive, and mixed respiratory pauses in both infant groups.
The referral infants had a lower baseline Sao2 (95 ± 1% vs 98 ± 1%, P < .003) than the controls. However, the lowest saturation associated with an event was comparable between groups (79 ± 1% vs 81 ± 1%, referral vs control infants, respectively). Neither time from onset of respiratory pause to onset of desaturation (10.0 ± 1.0 vs 10.1 ± 0.5 seconds) nor time to lowest saturation (17.0 ± 1.0 vs 18.9 ± 0.7 seconds, referral vs control infants, respectively) differed between infant groups (Fig 1A).
|
80%; however, 94% of the bradycardia events had a decrease in Sao2 of at least 5%. The referral infants had a higher number of periodic breathing cycles (224 ± 35 vs 133 ± 20 cycles/12-hour study, P < .03). These cycles of periodic breathing were associated with 409 of 839 desaturation or bradycardia events in the referral infants versus 66 of 219 in the control infants (P < .001). However, the increased occurrence of periodic breathing did not account for all the differences between infant groups because the remaining number of desaturation or bradycardia events associated with isolated respiratory pauses was also higher in the referral infants (430/839 vs 153/219 desaturation/bradycardia events, P < .001). The respiratory pauses associated with bradycardia and desaturation that occurred during isolated apnea were of comparable duration to those that occurred during periodic breathing.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
20 seconds, with the frequency of occurrence being comparable to that of infants having no clinical bedside monitor alarms. However, these 2 infant groups could be clearly differentiated by the higher incidence of episodic desaturation and bradycardia associated with shorter respiratory pauses and a lower baseline Sao2 in the referral infants. Poets et al13 have shown that apparently well preterm infants exhibit periods of hypoxemia associated with short respiratory pauses not detected by routine clinical bedside monitoring. This study has confirmed those findings in the control infants. In addition, we have shown that infants referred for apnea monitoring studies have a 3 times more such desaturation events than control infants.
The higher occurrence of desaturation episodes in the referral infants may be partly attributed to differences in baseline Sao2 measurements between groups. Previously published Sao2 values have ranged from 91% to 100% (median 99.4%) for preterm infants breathing room air.14 Baseline measurements of Sao2 were within this range for both infant groups; however, the median Sao2 level for control infants (98.4%) was similar to previously published values of 99.4%, whereas the median Sao2 for referral infants (95.5%) fell below this value. The lower baseline Sao2 for referral infants would be expected to make them more vulnerable to desaturation in response to a respiratory pause, although the decrease in Sao2 was neither faster nor lower in the referral group. Although all infants were asymptomatic at the time of study and none were receiving supplemental oxygen, the lower baseline Sao2 may indicate that the referral infants had more residual lung disease, with a trend toward a greater number of days on supplemental oxygen or mechanical ventilation (Table 1). These differences did not reach statistical significance, which may have resulted from an inadequate sample size for these 2 variables. Future studies might evaluate the optimal baseline Sao2 for such infants.
As with desaturation, the increased occurrence of bradycardia episodes in the referral infants may be attributable to lower baseline levels of Sao2. This is supported by previous findings that apneic episodes accompanied by bradycardia are associated with a lower baseline Sao2 than apnea without accompanying bradycardia.15 Alternatively, the higher incidence of bradycardia episodes may be attributed to increased vagal tone in the referral group. This may account for the shorter time from onset of the respiratory pause to the initial decrease in heart rate and to the lowest heart rate during the event. Finer et al16 have shown that bradycardia occurs more often with central than with mixed or obstructive apnea, suggesting that the heart rate response depends on the type of apnea. However, in this study bradycardia episodes were evenly distributed between central and mixed respiratory pauses in both infant groups.
It has been hypothesized that bradycardia occurs as a reflex response to the decrease in Sao2 induced by apnea.17 However, studies comparing the relationship between the initial decrease in Sao2 and heart rate reveal inconsistent results.4,5 This study showed that the initial decrease in heart rate preceded the decrease in Sao2 in both groups of infants. However, because of the Sao2 algorithm that averages the pulse wave form over a 3-second window and the effect of the infant’s circulation time, it is difficult to compare timing relationships between averaged Sao2 and beat-to-beat changes in heart rate. Previous data by Poets et al18 have shown that 86% of documented bradycardia was associated with desaturation 80%. Our study demonstrated that 18% of the bradycardia events were associated with desaturation, defined as 80%; however, 94% of the bradycardia events were associated with a decrease in Sao2 of at least 5%. Furthermore, desaturation and bradycardia were invariably preceded by a respiratory pause. Therefore, although the precise relationship between desaturation and bradycardia remains unclear, episodic apnea, desaturation, and bradycardia are clearly linked.
The mean occurrence of apnea 20 seconds and the proportion of infants having at least 1 apnea 20 seconds in duration were higher in the referral infants, but neither comparison approached statistical significance. This may have been because of an inadequate sample size. To detect a difference of 12% for the controls versus 23% for the referrals in having at least 1 apnea 20 seconds, as found in this study, a sample of 187 infants for each group would be needed with a power of 0.8 and 
= .05.
This study did not seek to characterize causality for cardiorespiratory events, so a case-controlled or matched study design was not used. Instead, the study was designed to record and describe events in referral and control infants of similar postconceptional age. To avoid large-group incomparability, enrollment criteria entailed a group-matched design of postconceptional ages between 33 and 36 weeks for both control and referral infants. All infants fell within this age range, but after the target sample size was reached the referral infants were statistically older by 0.8 weeks’ postconceptional age at the time of study. This should not have affected our conclusion that the referral infants were more vulnerable to episodic desaturation because the higher postconceptional age should make this group less vulnerable to respiratory instability.
The referral infants were less mature at birth, as defined by a lower gestational age, a parameter we did not seek to control for, and this may have contributed to the enhanced vulnerability of this group. In contrast, the referral infants had a higher postnatal (and postconceptional) age, indicative of greater maturity at the time of the study. It is unclear how these discrepancies in ages may have affected maturation of cardiorespiratory control.
A limiting factor of this study is the absence of a standard clinical protocol to guide attending physicians in referring infants for overnight apnea monitoring studies at Rainbow Babies and Children’s Hospital or at most other institutions. Therefore, referral of infants for this study depended solely on a request for consultation from individual attending staff, and this may have affected the infants included in the referral group. However, referral infants were enrolled over a period of 2 years, with an average of 3 infants enrolled per month from a group of 14 attending physicians, with no individual attending dominating enrollment. It should be noted that monitor alarm settings did not vary over time or between attending physicians.
Control infants were defined as having no events for 2 days before the study, based on clinical criteria including bedside monitoring and charting of such events. It is impossible to say with total confidence that the infants were entirely event-free over this period. However, if controls were not event-free, this should have diminished any differences in cardiorespiratory events between groups.
We conclude that infants referred for apnea monitoring studies because of persistent bedside monitor alarms have very infrequent apnea of at least 20 seconds’ duration but a high frequency of desaturation and bradycardia in response to short respiratory pauses when compared with infants with no persistent bedside monitor alarms. Such referral infants also have a lower baseline Sao2, which may indicate subclinical residual lung disease. Chronic lung disease, as defined by a persistent oxygen requirement, has been associated with a delay in resolution of apnea or bradycardia3 and poor neurodevelopmental outcome.19,20,21 Future studies must consider the role of abnormal cardiorespiratory control and recurrent desaturation as markers of the impaired neurodevelopmental outcome that often accompanies chronic lung disease in the low birth weight survivors of neonatal intensive care.
|
FOOTNOTES |
|---|
Reprint requests to (J.M.D.) Division of Neonatology, Rainbow Babies and Children’s Hospital, 11100 Euclid Ave, Cleveland, OH 44106. E-mail: jmd3{at}po.cwru.edu
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
 |
|
 |
|
  J. M. Di Fiore, M. Arko, M. Whitehouse, A. Kimball, and R. J. Martin Apnea Is Not Prolonged by Acid Gastroesophageal Reflux in Preterm Infants Pediatrics, November 1, 2005; 116(5): 1059 - 1063. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. M. Baird, R. J. Martin, and J. M. Abu-Shaweesh Clinical Associations, Treatment, and Outcome of Apnea of Prematurity NeoReviews, April 1, 2002; 3(4): e66 - 70. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
||
|
||
Home | My Pediatrics | Journal Information | Current Issue | Past Issues | Subscriptions & Services | Contact Us Click here for faster international access © 2001 American Academy of Pediatrics. All rights reserved. |
||
|
||
) had both an earlier onset of heart rate deceleration and shorter time to lowest heart rate after a respiratory pause than the control infants (
).