Etiology and Mechanism of Intermittent Hypoxemia Episodes in Spontaneously Breathing Extremely Premature Infants

Abstract

Objective:

To evaluate the mechanisms leading to intermittent hypoxemia (IH) episodes in spontaneously breathing extremely premature infants at 32 and 36 weeks postmenstrual age (PMA).

Methods:

We studied spontaneously breathing premature infants of 23–28 weeks gestational age who presented with IH episodes while on non-invasive respiratory support at 32 or 36 weeks post menstrual age (PMA). Daytime recordings of arterial oxygen saturation (SpO₂), esophageal pressure, respiratory inductive plethysmography of the abdomen, chest wall and their sum were obtained during 4 hours at 32 and 36 w PMA. IH episodes (SpO₂ < 90% for ≥ 5 s) and severe IH episodes (SpO₂ < 80% for ≥ 5 s) were classified as resulting from apnea, active exhalation and breath holding, reduced tidal volume (VT) or reduced RR during the preceding 60 s.

Results:

Fifty-one infants of (mean ± SD) 25.9 ± 1.5 weeks gestational age and 846 ± 185 g birth weight were included, of these, 31 and 41 were included in the analysis at 32 and 36 weeks PMA, respectively. At both 32 and 36 weeks PMA a greater proportion of all IH episodes and severe IH episodes were associated with active exhalation and breath holding than with apnea, reduced RR or reduced VT. The severity and duration of the IH episodes did not differ between mechanisms.

Conclusions:

In this group of premature infants, the predominant mechanism associated with daytime IH was active exhalation and breath holding. This etiology is more associated with behavioral factors than abnormal respiratory control and can have implications for prevention.

Premature infants often present with oxygenation instability manifested as episodes of intermittent hypoxemia (IH). These episodes vary in frequency and may last from several seconds to minutes. IH episodes have been associated with increased risk for neurodevelopmental impairment, severe retinopathy of prematurity (ROP) and bronchopulmonary dysplasia [1–4].

In spontaneously breathing preterm infants IH episodes are generally attributed to central apnea as the predominant mechanism [5–7]. IH episodes are also observed in preterm infants receiving mechanical ventilation. These are often triggered by contractions of the abdominal musculature that increase abdominal and intrathoracic pressure, producing active exhalation that reduces lung volume, and breath holding that impairs ventilation despite continuous cycling of the ventilator [8]. It has also been shown that consecutive abdominal contractions with the resulting decrease in FRC and ventilation are associated with increased severity and duration of IH episodes [9]. This mechanism also contributes to the occurrence of IH episodes in spontaneously breathing preterm infants after weaning from mechanical ventilation. In these infants, the majority of IH episodes were not associated with central apnea but instead with contraction of the abdominal muscles [10]. However, it is unknown if the same mechanism contributes to oxygenation instability and IH in spontaneously breathing infants at later postnatal ages and at near term corrected age.

The objective of this study was to determine the prevalence of the different mechanisms that lead to intermittent hypoxemia episodes in spontaneously breathing premature infants at 32 and 36 weeks postmenstrual age (PMA).

Methods

This is a secondary analysis of data collected by the prospective observational study conducted at the Newborn ICU at Holtz Children’s Hospital of the University of Miami/Jackson Memorial Medical Center as part of the Prematurity Related Ventilatory Control (Pre-Vent) study sponsored by the National Heart Lung and Blood Institute of the National Institutes of Health at this site [11]. Preterm infants of 23^0/7– 28^6/7 weeks gestational age (GA) who received supplemental oxygen, invasive mechanical ventilation, or non-invasive respiratory support at any time during their hospitalization were included in the study. Infants with severe congenital anomalies that may affect life expectancy or pulmonary or neurosensory development and severe CNS pathology that may alter respiratory control function were excluded. Infants were enrolled after written informed parental consent. The study was conducted under approval of the University of Miami Institutional Review Board and the Jackson Health System Clinical Research Office and it was overseen by an independent observational safety monitoring board appointed by the sponsor.

A sub-cohort of spontaneously breathing infants receiving nasal CPAP, non-invasive ventilation (NIV), nasal cannula flow (NC) or on room air who had IH episodes at 32 or 36 weeks PMA were included in the analysis for this ancillary study. Infants were not studied when vaccinations or ophthalmology exams had been administered in the preceding two days.

Day time recordings of arterial oxygen saturation (SpO₂), esophageal pressure (P_ESOPH), respiratory inductive plethysmography (RIP) of the abdominal (RIP_ABD), chest wall (RIP_CW) and the sum of the two RIP bands (RIP_SUM) were obtained for 4 hours at 32 and 36 weeks PMA. SpO₂ was measured using a neonatal pulse oximeter (Radical 7, Masimo, Irvine, CA) with 8 s averaging time. P_ESOPH was measured by a water filled 3.5Fr catheter with the tip at the lower third of the esophagus and connected to a pressure transducer (Deltran, Utah Medical, Salt Lake City, UT). RIP was measured by a respiratory inductive plethysmograph (Respitrace, Sensormedics, Yorba Linda, CA). RIP signals were calibrated using the built-in method of qualitative diagnostic calibration but were not calibrated to tidal volume. All analog output signals were digitized and recorded into a computerized system (Windaq Pro/DI200, Dataq Instruments, Akron, OH) at 100 samples/s.

For purposes of this study, the definition of IH episodes was a decline in SpO₂ below 90% for ≥ 5 s. Consecutive events with SpO₂ declining below 90% occurring less than 5 seconds apart were considered as a single episode. The basal SpO₂ was determined during the 30 s before the decline in SpO₂. The duration of the episode was determined as the time SpO₂ remained below 90%. The nadir of each episode was defined as the lowest SpO₂ during the episode. Severe IH episodes were those with SpO₂ declining below 80% for ≥ 5 s. The duration of the severe IH episode was determined as the time SpO₂ remained below 80%. Severe IH episodes were a subset of IH episodes below 90%.

To determine the mechanism associated with each IH episode, RIP and P_ESOPH waveforms during the 60 seconds before the start of the IH episode were examined. IH episodes were classified as those resulting from the following mechanisms: Central apnea, active exhalation and breath holding, reduced respiratory rate (RR) or reduced tidal volume (VT) during the preceding 60 seconds. The definition of central apnea for purposes of this study was a breathing pause ≥ 5 s with absence of inspiratory deflections in the P_ESOPH waveform and VT in the RIP waveforms. The working definition for the mechanism described as active exhalation and breath holding was an increase in intrathoracic pressure identified as a rise in P_ESOPH that produced an active exhalation resulting in a decrease in lung volume shown by a decrease in baseline of the RIP_ABD and/or RIP_CW waveforms accompanied by breath holding shown by reduced or disorganized ventilation in the RIP waveforms. Reduced RR was defined as a decline in RR noted in the RIP_ABD, RIP_CW and P_ESOPH waveforms that could not be classified as central apnea or as the active exhalation and breath holding mechanism. Reduced VT was defined as a decline in the amplitude of the RIP_SUM and P_ESOPH waveforms that could not be classified as central apnea or as the active exhalation and breath holding mechanism. Episodes were also classified as those resulting from active exhalation and breath holding preceded or followed by central apnea. For each infant, IH episodes classified according to the associated mechanism were counted and the percentage out of all IH episodes was calculated. Fluctuations in SpO₂ during segments of periodic breathing (PB) were not included in the analysis. For this study a segment of PB was defined as a consistent rhythm of more than 3 cycles with less than 10 seconds of breathing and pauses of at least 3 seconds between cycles.

During the bedside studies the center’s respiratory guidelines were followed. SpO₂ alarms were set at 90 and 95% and FiO₂ was adjusted to keep SpO₂ within 90–95%. In the event SpO₂ decreased below 90% FiO₂ was increased by .05–0.10 within 30 seconds and the infant was stimulated in case of apnea. If SpO₂ decreased below 80% FiO₂ was increased by 0.10–0.20 within 30 seconds. These steps were repeated if SpO₂ remained below the target range. FiO₂ was brought back to the preceding baseline level at the end of the event. Placement of the nasal interface and airway patency were monitored during the study.

Statistical analysis consisted of within-subject comparisons using Linear Mixed Models (LMM) with repeated measures and post-hoc paired comparisons by the Sidak method using the IBM SPSS Statistics package Version 27 (International Business Machines Corporation, Armonk, NY). A p < .05 was considered statistically significant.

Results

Fifty-four infants were enrolled from October 2018 to May 2022 and underwent study procedures. Fifty-one of these infants were included in the analysis because they presented with IH episodes while breathing spontaneously on non-invasive respiratory support at 32 weeks or 36 weeks PMA, or both. Mean gestational age (± SD) was 25.9 ± 1.5 weeks and mean birth weight was 846 ± 185 grams. Twenty-five of infants (48%) were female, 26 (50%) were black and 23 (44%) were Hispanic.

Of the 51 infants included in the analysis, 31 and 41 infants were included at the 32 weeks and 36 weeks PMA time points, respectively; 21 infants were included in the analysis at both 32 weeks and 36 weeks PMA time points. Fewer infants were included at 32 weeks PMA because some of them were receiving invasive mechanical ventilation or because enrollment occurred after 32 w PMA.

Figure 1 shows representative tracings during an episode of IH preceded by a brief apnea. In these tracings apnea is noted by the absence of spontaneous inspiratory breathing activity and tidal expansion in the RIP waveforms that is followed by a decline in SpO₂. Figure 2 shows representative examples of IH episodes associated with active exhalation and breath holding. The figure shows the increase in intrathoracic pressure that is transmitted to the esophagus and produces an active exhalation with loss in lung volume. This is followed by increases in intrathoracic pressure that produce a breath holding pattern that impairs ventilation. The second episode illustrates the same mechanism that is preceded by a brief apnea. Rapid and vigorous breathing efforts are present at the end of the IH episodes.

Figure 1. Representative example of episodes of hypoxemia associated with apnea.

Figure shows representative tracings during an episode of hypoxemia preceded by a brief apnea. The vertical line marks the start of the decline in arterial oxygen saturation (SpO₂, top panel) below 90% (dashed horizontal line). Apnea is noted by the absence of spontaneous breathing efforts (no respiratory activity noted in the esophageal pressure waveform) that is accompanied by absence of tidal expansion in the respiratory inductance plethysmograph (RIP) waveforms of the bands at the chest, abdomen and their sum.

Figure 2. Representative examples of hypoxemia episodes associated with active exhalation accompanied by breath holding secondary to due to increases in intrathoracic pressure.

Figure shows representative tracings obtained during two consecutive episodes of hypoxemia associated with active exhalation and breath holding. The vertical line marks the start of the decline in arterial oxygen saturation (SpO₂, top panel) below 90% (dashed horizontal line) in the first episode. This episode is preceded by increases in intrathoracic pressure that are transmitted to the esophagus. The active exhalation resulted in loss in lung volume noted as a decline in the baseline of the respiratory inductance plethysmograph (RIP) waveforms. Subsequent increases in intrathoracic pressure produced a breath holding pattern that impaired ventilation as shown in the RIP waveforms. The second episode of hypoxemia was preceded by a brief apnea, illustrated by the absence of inspiratory deflections in the esophageal pressure waveform, that was followed by acute increases in intrathoracic pressure that produced active exhalation and periods of disrupted ventilation. There were some relatively large inspiratory efforts observed in the esophageal pressure waveform interspersed between the acute increases in pressure. The resolution of the IH episodes was started by rapid and vigorous breathing effort that increased ventilation above the basal levels prior to the episodes.

Figures 3 and 4 show representative examples of IH episodes associated with a reduction in tidal volume from a declining inspiratory effort and a reduction in RR, respectively. The proportions of IH episodes associated with these two mechanisms were relatively low.

Figure 3. Representative example of hypoxemia episodes associated with brief apnea and with a reduction in tidal volume.

Figure shows representative tracings of two mild IH episodes with SpO₂ declining just below 90% (top panel). The first is a brief episode associated with a short breathing pause. The second episode lasted approximately 17.5 seconds (marked by the two vertical lines) and was associated with a reduction in tidal volume from a gradual decline in inspiratory effort (smaller negative deflection in the esophageal pressure waveform).

Figure 4. Representative example of a hypoxemia episodes associated with a reduction in respiratory rate.

Figure includes representative recordings during a mild IH episode with SpO₂ declining just below 90% (top panel) for approximately 21.5 seconds (marked by the two vertical lines). A reduction in respiratory rate (RR) precedes the decline in SpO₂. The reduction in RR is observed in the esophageal pressure and RIP waveforms. The resolution of the IH episode was accompanied by rapid breathing effort.

Analysis at 32 weeks PMA:

Of the 31 infants included in the analysis at 32 w PMA, 3 infants were on nasal CPAP, 1 infant on NIV and 27 infants on NC. Thirteen of the 31 infants were on supplemental oxygen with a mean FiO₂ of .27 ± .05 and 28 infants (90%) were receiving caffeine. Their postnatal age was 38.7 ± 11.1 days at the time of study. Table 1 shows the number, proportion, preceding baseline level, duration and nadir level of the IH episodes (SpO₂ < 90% for ≥ 5 s) according to their associated mechanism at 32 w PMA. A greater proportion of IH episodes were associated with active exhalation and breath holding than with central apnea, or reduced RR or VT. Table 2 shows the number, proportion, preceding baseline level, duration and nadir of the severe IH episodes (SpO₂ < 80% for ≥ 5 s). Similarly, a greater proportion these episodes were associated with active exhalation and breath holding than with central apnea, or reduced RR or VT.

Table 1.

Episodes of intermittent hypoxemia classified according to trigger mechanism at 32 w PMA

	Central apnea	Active exhalation and breath holding	Active exhalation and breath holding preceded by central apnea	Active exhalation and breath holding followed by central apnea	Reduced RR	Reduced VT
Number of IH episodes / 4 hours	2.4 ± 4.6	13.1 ± 11.7a	1.0 ± 1.7	0.9 ± 1.4	1.7 ± 2.4	1.1 ± 1.8
Proportion of IH episodes (%)	10.5 ± 15.0	64.4 ± 25.3a	5.4 ± 9.6	4.8 ± 9.9	9.5 ± 14.4	5.5 ± 8.9
Basal SpO2 before IH episode (%)	93.8 ± 2.5	93.3 ± 1.9	93.9 ± 2.7	93.1 ± 1.7	92.6 ± 2.4	92.4 ± 1.6
IH episode duration with SpO2 < 90% (s)	32.6 ± 31.1	56.7 ± 72.7	57.6 ± 59.8	63.8 ± 46.1	79.9 ± 61.7	57.9 ± 72.7
Nadir SpO2 during IH episode (%)	81.3 ± 4.6	82.2 ± 2.7b	80.7 ± 3.2	77.6 ± 7.0	81.7 ± 4.9	83.0 ± 4.1b

IH episodes defined as a decline in SpO₂ below 90% for at least 5 seconds.

Data are reported as Mean ± SD

^ap < .001 vs all others

^bp < .05 vs active exhalation and breath holding followed by central apnea

Table 2.

Episodes of severe intermittent hypoxemia classified according to trigger mechanism at 32 w PMA

	Central apnea	Active exhalation and breath holding	Active exhalation and breath holding preceded by central apnea	Active exhalation and breath holding followed by central apnea	Reduced RR	Reduced VT
Number of IH episodes / 4 hours	.58 ± 1.0	4.0 ± 5.6a	.46 ± .91	.58 ± 1.1	.62 ± 1.5	.31 ± .74
Proportion of IH episodes (%)	11.6 ± 21.8	55.3 ± 35.7a	6.8 ± 12.7	6.5 ± 13.1	12.2 ± 24.7	7.7 ± 21.8
Basal SpO2 before IH episode (%)	93.4 ± 1.9	92.8 ± 2.3	92.8 ± 1.5	93.3 ± 1.9	93.5 ± 3.3	91.3 ± 1.3
IH episode duration with SpO2 < 80% (s)	17.2 ± 13.3	14.8 ± 6.2	23.7 ± 24.0	21.0 ± 6.2	20.8 ± 9.8	18.9 ± 14.8
Nadir SpO2 during IH episode (%)	72.7 ± 6.1	76.0 ± 2.7	76.4 ± 3.4	72.3 ± 6.2	73.1 ± 5.9	76.1 ± 2.3

IH episodes defined as a decline in SpO₂ below 80% for at least 5 seconds.

Data are reported as Mean ± SD

^ap < .001 vs all others

The basal SpO₂ preceding the episode, the episode severity as indicated by the nadir level reached during the episode and the episode duration did not differ between episodes associated with the different mechanisms.

Analysis at 36 weeks PMA:

Of the 41 infants included in the analysis at 36 w PMA, 1 infant was on nasal CPAP, 1 on NIV and 39 infants on NC. Eleven of the 41 infants were receiving supplemental oxygen with a mean FiO₂ of .29 ± .08 and 27 infants (66%) were receiving caffeine. Their postnatal age was 70.1 ± 11.3 days at the time of study. Table 3 shows the number, proportion, preceding baseline level, duration and nadir of the IH episodes (SpO₂ < 90% for ≥ 5 s) according to their associated mechanism at 36 w PMA. At this age a greater proportion of IH episodes were associated with active exhalation and breath holding in comparison to the other mechanisms. The severity or duration of the episodes did not differ between IH episodes associated with the different mechanisms. Table 4 shows the number, preceding baseline level, duration and nadir of the severe IH episodes (SpO₂ < 80% for ≥ 5 s). A greater proportion of IH episodes were associated with active exhalation and breath holding when compared with the other mechanisms. The severity or duration of the episodes did not differ between mechanisms.

Table 3.

Episodes of intermittent hypoxemia according to triggering mechanism at 36 weeks PMA

	Central apnea	Active exhalation and breath holding	Active exhalation and breath holding preceded by central apnea	Active exhalation and breath holding followed by central apnea	Reduced RR	Reduced VT
Number of IH episodes / 4 hours	1.2 ± 2.3	14.0 ± 10.6a	1.1 ± 2.0	0.4 ± 0.9	1.4 ± 2.9	0.3 ± 0.9
Proportion of IH episodes (%)	4.0 ± 8.1	81.6 ± 20.1a	3.5 ± 5.9	1.8 ± 3.7	8.0 ± 13.6	0.9 ± 2.4
Basal SpO2 before IH episode (%)	9.4 ± 2.5	93.2 ± 1.9	93.9 ± 2.7	93.1 ± 1.7	92.6 ± 2.4	92.3 ± 1.5
IH episode duration with SpO2 < 90% (s)	26.4 ± 14.8	50.4 ± 26.4	44.1 ± 27.5	27.3 ± 12.4	45.9 ± 37.2	26.4 ± 12.7
Nadir SpO2 during IH episode (%)	80.5 ± 5.6	80.7 ± 3.7	79.4 ± 4.2	75.6 ± 13.0	80.6 ± 5.0	78.8 ± 5.7

IH episodes defined as a decline in SpO₂ below 90% for at least 5 seconds.

Data are reported as mean ± SD

^ap < .001 vs all others

Table 4.

Episodes of severe intermittent hypoxemia classified according to trigger mechanism at 36 w PMA

	Central apnea	Active exhalation and breath holding	Active exhalation and breath holding preceded by central apnea	Active exhalation and breath holding followed by central apnea	Reduced RR	Reduced VT
Number of IH episodes / 4 hours	.44 ± .97	4.9 ± 4.8a	.61 ± 1.4	.11 ± .32	.42 ± 1.4	.11 ± .39
Proportion of IH episodes (%)	7.3 ± 20.2	79.9 ± 29.2a	5.6 ± 13.3	.92 ± 3.5	5.4 ± 13.3	.80 ± 2.7
Basal SpO2 before IH episode (%)	94.1 ± 2.2	94.2 ± 2.0	94.5 ± 1.8	96.3 ± 3.9	92.7 ± 2.9	92.6 ± 3.0
IH episode duration with SpO2 < 80% (s)	8.3 ± 3.2	25.8 ± 23.9	15.4 ± 11.2	29.3 ± 22.3	16.6 ± 18.3	11.8 ± 6.8
Nadir SpO2 during IH episode (%)	74.9 ± 4.6	74.9 ± 3.3	75.4 ± 1.9	60.5 ± 13.2b	74.1 ± 6.7	71.8 ± 3.9

IH episodes defined as a decline in SpO₂ below 80% for at least 5 seconds.

Data are reported as Mean ± SD

^ap < .001 vs all others

^bp < .05 vs all others

Discussion

The data obtained during daytime in this group of preterm infants presenting with episodes of intermittent hypoxemia while breathing spontaneously on non-invasive respiratory support, showed that the majority of IH episodes at 32 and 36 weeks PMA were associated with active exhalation accompanied by breath holding. This was true for all episodes with SpO₂ < 90% as well as for the most severe IH episodes with SpO₂ < 80%.

The events resulting in IH episodes were consistently marked by an initial rise in intrathoracic pressure that produced an active exhalation and resulted in a decrease in lung volume. This was followed by subsequent increases in intrathoracic pressure that resulted in a breath holding pattern that impaired ventilation. These events led to the subsequent decline in SpO₂.

Examination of the events occurring in IH episodes associated with this mechanism revealed some relatively large inspiratory efforts interspersed between acute increases in intrathoracic pressure, which indicates increased respiratory drive due to the lower arterial oxygen and higher CO₂ levels. The persistence of the increases in intrathoracic pressure that impair ventilation despite the higher respiratory drive is intriguing. The resolution of the IH episode generally started with of a period of rapid and vigorous breathing that was followed by an increase in SpO₂. The ventilation levels associated with the increased respiratory drive during the recovery phase were noticeably higher than the basal ventilation before the IH.

The proportion of IH episodes associated with this mechanism was considerably greater than those associated with central apnea or a spontaneous reduction in RR or VT. These findings suggest a new paradigm to the understanding of respiratory and oxygenation instability in premature infants. This mechanism for IH was previously described in mechanically ventilated infants as well as in infants receiving nasal CPAP after extubation [8–10, 12] but had not been described in spontaneously breathing premature infants at more advanced ages. Most IH episodes in spontaneously breathing infants had been previously attributed to central apnea [7, 13]. Interestingly, recordings obtained in this group of infants did not reveal IH events associated with obstructive apnea that would have been characterized by complete paradoxical movement between the abdominal and chest RIP signal with continued inspiratory efforts.

However, because the recordings did not include nasal flow the occurrence of obstructive apnea as a trigger mechanism for IH cannot be completely ruled out.

There may be multiple reasons for the observed difference in prevalence between central apnea and the mechanism of active exhalation and breath holding as triggers for IH episodes. Most infants included in this study were receiving caffeine which can effectively reduce central apnea in this population. This may have reduced the proportion of IH events triggered by apnea in this study. Also, infants who present with persistent apnea due to an immature respiratory control system may remain on invasive mechanical ventilation which made them ineligible for inclusion in this cohort of spontaneously breathing infants.

In this study measurements were obtained during daytime only. The observed difference in prevalence between mechanisms for IH episodes may be related to sleep state and behavioral factors during daytime. The prevalence of these mechanisms for IH may differ at night. It has been reported that agitation and increased activity are associated with more IH episodes in premature infants [12] and the frequency of IH episodes is higher during periods of wakefulness or indeterminate sleep state compared with quiet sleep [14]. These two sleep states account for most of the time in premature infants in the NICU. In mechanically ventilated infants, IH episodes have been shown to be more frequent during day compared with nighttime [15] which is likely related to the number of procedures and other activities in the NICU that can disrupt the preterm infant’s sleep and produce agitation. These observations suggest an important role of behavioral factors or other negative stimuli that increase patient activity and agitation that can trigger IH episodes.

Previous studies describing the mechanism of active exhalation and breath holding in mechanically ventilated infants and after weaning from mechanical ventilation measured abdominal electromyography and reported that the observed rise in abdominal and intrathoracic pressure was caused by contraction of the abdominal musculature. [9, 10] Although this study did not include recordings of abdominal electromyography, contraction of the abdominal muscles is the most likely spontaneous event that can explain the large increases in intrathoracic pressure observed in this study. Also, this study did not include video recordings which could have enabled analysis of the infant’s behavior around the time of the IH episodes.

All the infants included in this study had IH episodes but the frequency of the episodes was lower than that reported in mechanically ventilated infants [8, 9] and also below the reported frequency that peaks around postnatal week 4 [1, 16]. This difference is likely due to the relatively better lung function with only a small number of infants requiring non-invasive respiratory support and supplemental oxygen specially at 36w PMA.

The working definition of central apnea as a breathing pause of at least 5 seconds used in this study may be considered short compared with the standard definition of at least 20 seconds. However, this definition is consistent with the findings of a study that showed breathing pauses of 5–9 seconds can impact oxygenation in premature infants. [17] These relatively brief apnea events may have influenced the duration of the IH episodes which, although not statistically significant, were on average shorter than those caused by active exhalation and breath holding.

It can be argued that the working definition of IH episodes used in this study did not necessarily represent severe hypoxemia and included a number of mild episodes where SpO₂ did not drop below 80%. Nonetheless, these milder episodes may be relevant at near term corrected age and particularly when these infants approach the time for hospital discharge.

In this study RIP was used to measure the decrease in lung volume and ventilation and esophageal pressure was used to measure the increase in intrathoracic pressure associated with the active exhalation and breath holding events. Identifying this mechanism for IH in the NICU may not be straightforward because these methods are not commonly used for monitoring in the NICU. However, transthoracic impedance waveforms in bedside monitors may show decreases in lung volume depending on the specific monitor settings and show changes in the baseline of the impedance waveforms.

The pulse oximeter used in this study is reported to be resistant to motion artifact [18, 19] and has been used to detect IH and hyperoxemia in multiple recent observational and interventional studies in this population. A recent study evaluated the reliability of this pulse oximeter to detect hypoxemia in the presence of motion artifact by analyzing plethysmograph waveforms in a group of extremely preterm infants. [20] This study found SpO₂ measurements with this device in the presence of motion artifact were much more likely to show true- than false-hypoxemia.

The findings from this study suggest an etiology of IH episodes that is more influenced by behavioral factors than by abnormal control of breathing function. This provides new evidence of the factors involved in ventilatory and oxygenation instability in extremely premature infants and opens an opportunity for developing new and more effective strategies for prevention and management of these IH episodes.

In conclusion, the observations from this study indicate that the majority of daytime IH episodes in spontaneously breathing extremely preterm infants at near term corrected age were not due to central apnea, but instead they were preceded by active exhalation and breath holding produced by acute increases in intrathoracic pressure. These findings offer a new paradigm for the etiology of IH that seem to be more associated with behavioral factors than with abnormal respiratory control. We suggest that development of behavioral interventions should be explored as potentially more effective strategies to prevent or attenuate the severity of IH episodes.