Ventilatory Response of the Newborn Infant toMild Hypoxia

Gary Cohen, PhD, Girvan Malcolm, MB ChB, and David Henderson-Smart, PhD, FRACP*

Summary. The transition from an immature (biphasic) to a mature (sustained hyperpneic)response to a brief period of sustained hypoxia is believed to be well advanced by postnatal day10 for newborn infants. However, a review of the supporting evidence convinced us that thisissue warranted further, more systematic investigation. Seven healthy term infants aged 2 daysto 8 weeks were studied. The ventilatory response (VR) elicited by 5 min breathing of 15% O2

was measured during quiet sleep. Arterial SaO2(pulse oximeter) and minute ventilation (ex-

pressed as a change from control, DV8i) were measured continuously. Infants were wrapped intheir usual bedding and slept in open cots at room temperature (23°–25°).

Infants aged 2–3 days exhibited predominately a sustained hypopnea during the period ofhypoxia (DV8i = −2% at 1 min, −13% at 5 min). At 8 weeks of age, the mean response wastypically biphasic (DV8i = +9% at 1 min, −4% at 5 min). This age-related difference betweenresponses was statistically significant (two-way ANOVA by time and age-group; interactionP < 0.05). These data reveal that term infants studied under ambient conditions during definedquiet sleep may exhibit an immature VR to mild, sustained hypoxia for at least 2 months afterbirth. This suggests that postnatal development of the O2 chemoreflex is slower than previouslythought. Pediatr. Pulmonol. 1997; 24:163–172. © 1997 Wiley-Liss, Inc.

Key words: hypoxemia; infant; quiet sleep; oximeter; ventilation.

INTRODUCTION

An understanding of the mechanisms by which O2

homeostasis is maintained at maturity (in adults) hasbeen compiled over many years by analyzing responsesto a variety of experimental forcing functions. Our un-derstanding of the nature of this control system in earlylife—particularly around the time of birth—is more ru-dimentary. At this age, the likelihood that rapid and pro-found hypoxemia may develop is far greater, and effec-tive ventilatory and arousal responses to such challengesare thought to be critical for survival. Persisting imma-ture (‘‘fetal-like’’) or inappropriate chemoreflex-mediated responses to hypoxemia at this age may in-crease the risk of respiratory failure during sleep.1

Most of what is known of developmental changes inthe hypoxic responsiveness of infants is based on studiescomparing the response to so-called classical steady-statehypoxia at different ages. This test measures the venti-latory response elicited by a sudden, mild, sustained re-duction in theFI,O2

(12–15% for 3–5 min). The essentialfeatures of the ventilatory response to this stimulus wereinitially described more than 30 years ago, and virtually

nothing new has been added to our knowledge since.Full-term, newborn infants reportedly exhibit a biphasicresponse: within 1–2 min of exposure to this stimulus,V8iincreases by up to 13%, but by the end of 5 min venti-lation has returned to, or fallen below the pre-hypoxic(resting) value.2–4 With increasing postnatal age, the hy-perpneic phase is sustained for longer, and the overallresponse becomes more monophasic in appearance. Thehyperpneic phase of the adult response lasts for 10–20min before any decrease inV8i becomes evident.6

It is believed that the switch from the more immature(biphasic) to the mature, adult-like VRMH is well ad-

Department of Obstetrics and Gynaecology, King George V MemorialHospital and The University of Sydney, New South Wales, Australia.

Contract grant sponsor: National Health and Medical Research Coun-cil of Australia.

*Correspondence to: Prof. D.J. Henderson-Smart, NSW Center forPrenatal Health Services Research, QEII Institute for Mothers andInfants (DO2), University of Sydney 2006, Australia.

Received 27 January 1997; accepted 17 May 1997.

Pediatric Pulmonology 24:163–172 (1997)

Original Articles

© 1997 Wiley-Liss, Inc.

vanced, if not complete by about the 10th postnatal dayfor infants born at term4,7 and slightly later for infantsborn before term.8 In reviewing this data, it became ap-parent to us that certain aspects of the original method-ology needed to be clarified and revised, and the testsrepeated before the results could confidently be used todescribe the developmental changes in hypoxic respon-siveness that normally occur after birth. For instance,technical as well as ethical constraints precluded anydirect measure of the stimulus (other than theFI,O2

) dur-ing these early studies, making it impossible to relate thestimulus to the response accurately. Early efforts alsotended to focus on the responses of infants studied underartificially warm, or ‘‘thermoneutral’’ conditions (envi-ronmental temperature >32°C). Temperature may have aprofound influence on the ventilatory response elicitedby hypoxia.9,10 We were particularly interested in defin-ing the developmental pattern of the VRMH elicited un-der ambient conditions under which most healthy termbabies are reared. The long-held belief that major devel-opmental changes in O2 sensitivity occur within 2 or soweeks after birth has resulted in little work being under-taken to examine responses of infants beyond this period.We believed it would be useful to extend the scope ofthese developmental studies to include older infants. Fi-nally, previous studies were not behavioral state specific:we recognize today that specific sleep states may be as-sociated with selective depression (or excitation) of somereflex thresholds.

The aims of this study were, therefore, to re-investigate the postnatal development of the VRMH ofhealthy term infants, with particular attention to measur-ing the responses under ambient conditions—at the en-vironmental temperature at which infants are normallyraised in the nurseries of this hospital (23°–25°C) andduring electroencephalographically defined, naturally oc-curring QS. The stimulus administered was one used byprevious investigators (FI,O2

4 15% O2 for 5 min). Toevaluate developmental changes in the ventilatory re-

sponse, we compared responses elicited within the firstpostnatal week with those elicited 2 months after birth.To permit more accurate matching of stimulus and re-sponse, pulse oximetry was used to characterize the rateand depth of the hypoxemia experienced by each infantfollowing a step reduction inFI,O2

.

MATERIALS AND METHODS

Subjects

Five healthy term infants were recruited from the post-natal wards of King George V Hospital, and studied on(or about) day 2, day 7, and again at 8 weeks of age (oneinfant was lost to follow-up at 8 weeks). Two additionalinfants of staff members who volunteered to let theirinfants participate, were also studied at age 8 weeks.Mean gestation at birth of the infants studied was 40 ±0.3 weeks (range, 39–41 weeks) and mean birth weightwas 3,411 ± 163 g (range, 2,735–3,965 g). The studyprotocol was approved by the Ethics Review Committeeof this hospital, and all studies were undertaken with theinformed consent of the parents and the attending phy-sician.

Measurements

Continuous recordings were made of the EEG, EOG(from a piezoelectric crystal taped over the eyelid), SaO2

(N-100 pulse oximeter, Nellcor, Hayward, CA) used inthe 3 s response mode with the probe attached to the foot,heart rate (a tachograph was triggered from the pulsesignal of the oximeter), and end-tidal PCO2

(PetCO2; Eng-

strom-Eliza Duo, Gambro-Engstrom AB, Bromma, Swe-den). Airflow was measured using a miniaturized nasalmask-pneumotachograph, a detailed description of whichhas been provided.11 Subjects added to (during expira-tion) or subtracted from (inspiration) a continuous streamof gas that passed through the pneumotachograph assem-bly at an approximate flow of 3.5 Lz min−1. The pressuretransducer connected to the pneumotachograph (M 45-2,Validyne Engineering, Northridge, CA) was electricallybalanced to compensate for this background flow. Allsignals were recorded on a polygraph (Grass Instruments,Quincy, MA) and 14-channel magnetic tape. Airflow,PetCO2

andSaO2signals were electronically digitized and

stored for later analysis (IBM-compatible personal com-puter with an Analogue Devices RTI-815 A/D con-verter).VT was calculated from digital integration of theairflow signal, using calibration signals which were en-tered at the beginning and completion of each study.11

Studies were performed between mid-morning andmid-afternoon. After the attachment of recording elec-trodes and a scheduled feed, infants were wrapped intheir usual bedding and placed in a supine or slightly

Abbreviations

ANOVA Analysis of varianceEEG ElectroencephalogramEOG Electrooculogramf Breathing rateFI,O2

Fractional inspired concentration of O2

PACO2Alveolar partial pressure of CO2

PetCO2End-tidalPCO2

QS Quiet sleepREM Rapid eye movementSaO2

Arterial oxygen saturationVR Ventilatory responseVRMH Ventilatory response to mild hypoxiaV8i Minute ventilationVT/Ti Mean inspiratory flowVT Tidal volume

164 Cohen et al.

lateral position in open cots. The environmental tempera-ture was that normally maintained in this hospital’s spe-cial care nursery (23°–25°C). Skin (axillary) temperaturewas measured at the start and finish of each study. Alltests were carried out during state 1 of Prechtl12 or quietsleep as defined by Anders et al.13 A two-way tap at-tached to the gas cylinder outlets permitted the compo-sition of the gas mixture passing through the pneumo-tachograph assembly to be changed rapidly from air to15% O2 (balance N2). All gases were delivered at roomtemperature and humidity. The test protocol was as fol-lows: after at least 3 min of unequivocal quiet sleep, datawere collected for 1 min to determine resting or controlvalues of respiratory parameters. The inspired gas wasthen suddenly switched to 15% O2 and maintained at thislevel for 5 min. Replicate responses were obtained when-ever possible. A minimum recovery period of 5 min waspermitted between tests, although usually only one testwas administered during each epoch of quiet sleep.

Data Analysis

Tests were not analyzed if there was a change of be-havioral state during the test (a reduction in voltage, orwaning of the trace-alternant or high-voltage pattern ofthe EEG), or if behavioral arousal (agitation, crying, eyeopening, and body movement) prevented the collectionof a minimum of 3 min of dataduring the period ofhypoxia.Breath-to-breath values ofV8i, VT, and f wereexpressed as percentage change from the mean of allbreaths recorded during the 1 min control period of eachtest; augmented breaths, or ‘‘sighs’’ (breaths with aVT >2× theVT of preceding breaths) were excluded from thedata because theVT of these breaths was not within therange over which the pneumotachograph was routinelycalibrated. MeanV8i, VT, f, VT/TI, PetCO2

, andSaO2were

calculated from all breaths recorded during consecutive30 s epochs for the duration of each hypoxic challenge. Ifsatisfactory duplicate test results were obtained during astudy, data points for each test were converted to per-centage changes from the respective control values and

combined, and an overall response curve was derived.Statistical comparison of means was by ANOVA withStudent-Newman-Keuls correction for multiple rangetests. Data, unless otherwise specified, are presented asthe mean ± SEM.

To facilitate objective analysis of the VRMH, we ap-plied arbitrary limits to the qualitative descriptions ofresponses reported in earlier studies, using the followingcriteria: V8i decreased:no significant increase inV8i inresponse to 15% O2, with V8i below the control level atthe end of 5 min and less than the controlV8i during morethan half of the 30 s epochs analyzed;V8i unchanged:noincrease inV8i during the test, withV8i not different fromcontrol at the termination of the test as well as duringmore than half the epochs analyzed;V8i increased: V8iincreased within 60 s, with an increase still evident at theend of the test as well as during more than half the epochsanalysed;biphasic response: V8i increased within 60 s,but not different from, or below the controlV8i from the3rd min onwards;periodic breathing:three or more epi-sodes of apnea of at least 2 s duration separated by regu-lar respirations occurring within the first 3 min of a test.14

Ventilatory parameters were not analyzed if periodicbreathing occurred because of the difficulty in ensuringaccurate measurement of theVT.

RESULTS

Seven infants were successfully studied on 16 occa-sions. Five infants were studied twice during the firstpostnatal week. Four of these infants, as well as twoadditional infants, were studied at age 8 weeks. Therewere 37 attempts to measure the VRMH, of which 26responses were considered suitable for analysis; 11 re-sponses were excluded for the reasons given above. Con-trol data (nasal mask attached, breathing air) for each ofthe respiratory variables are given in Table 1. Body tem-perature did not change significantly during the course ofthe studies (36.84 0.1°C at the start, versus 36.6 ±0.2°C at the conclusion of study;P > 0.3).

TABLE 1—Control Measurements 1

Age(days)

V8i(ml z kg−1 z min−1)

VT

(ml z kg−1)f

(breathsz min−1)PetCO2

(Torr)SaO2

(%)

2.5 ± 0.3 220 ± 21 5.1 ± 0.3 44 ± 4 35.4 ± 1.2 98 ± 1.0(161–366) (4.0–7.3) (25–71) (29.1–42.6) (95–100)

7.5 ± 0.7 207 ± 19 4.6 ± 0.3 46 ± 5 40.5 ± 1.0 98 ± 0.5(150–297) (3.5–5.7) (29–70) (37.1–43.0) (96–100)

56 ± 0 214 ± 17 5.2 ± 0.3 43 ± 6 39.2 ± 0.9 99 ± 0.3(144–293) (3.4–6.3) (32–80) (35.1–42.2) (98–100)

1Minute ventilation (V8i), tidal volume (VT), breathing rate (f), end-tidalPCO2(PetCO2

), and oxygen saturation (SaO2) of newborn infants breathing

air during the 60 sec immediately preceding the onset of a 5-min period breathing 15% O2. A nasal mask was used to measure airflow; air waspassed through this assembly at 3 Lz min−1. Data represent the mean ± SEM at three ages (first column); the range is indicated in parentheses.For details of individual responses, see Table 2.

Response of the Newborn to Hypoxia 165

TABLE 2—Ventilatory Response to 15% O 21

Subject(no. of tests) Age

V8i during inhalation of 15% O2 (% change from control; mean ± sem) Ventilatoryresponse309 609 909 1209 1509 1809 2109 2409 2709 3009

MCD (2) 3 days −5 ± 2 +5 ± 4 −8 ± 2 −13 ± 3* −15 ± 2* −16 ± 2* −15 ± 2* −12 ± 4* −19 ± 4* −14 ± 3* V8i DecreasedTUR (2) 3 days 0 ± 2 −3 ± 2 −7 ± 3 −10 ± 2* −14 ± 3* −15 ± 3* −15 ± 3* −6 ± 5 −14 ± 3* −18 ± 3* V8i DecreasedHIL (2) 2 days 0 ± 1 −4 ± 3 −3 ± 2 −11 ± 1* −13 ± 1* −14 ± 3* −14 ± 2* −17 ± 2* +6 ± 8 −11 ± 1* V8i DecreasedRED (2) 1.5 days −10 ± 2* −12 ± 1* −12 ± 2* −19 ± 2* −16 ± 2* −12 ± 2* −13 ± 2* −14 ± 2* −9 ± 3* −15 ± 2* V8i DecreasedWHE (2) 3 days +12 ± 2* +4 ± 1 −2 ± 2 −8 ± 2* −7 ± 3* −13 ± 2* −12 ± 2* −12 ± 2* −14 ± 2* −9 ± 4* V8i BiphasicMCD (2) 7 days +10 ± 3 +9 ± 2 +8 ± 3 +10 ± 4 +3 ± 3 +3 ± 5 +2 ± 7 −4 ± 6 −4 ± 8 +11 ± 13 V8i UnchangedHIL (2) 6 days +2 ± 4 −13 ± 3* −11 ± 3 −3 ± 1 −3 ± 2 −1 ± 1 −5 ± 2 +7 ± 7 −9 ± 3 −2 ± 2 V8i UnchangedRED (1) 8 days +11 ± 3 +5 ± 4 +23 ± 6* +16 ± 6 +14 ± 9 +6 ± 2 8 ± 1 +12 ± 4 +14 ± 4 +17 ± 3 V8i BiphasicMCD (2)2 8 weeks +6 ± 2* +10 ± 1* +10 ± 2* +5 ± 2* +1 ± 2 −3 ± 2 −1 ± 3 −1 ± 1 — — V8i BiphasicRED (1) 8 weeks +12 ± 4 +15 ± 4* +1 ± 8 +1 ± 5 +8 ± 7 +8 ± 6 +7 ± 4 +7 ± 3 +4 ± 4 +1 ± 5 V8i BiphasicWHE (1) 8 weeks +13 ± 3* +6 ± 2 +1 ± 2 +6 ± 9 −4 ± 5 −2 ± 3 −5 ± 2 −4 ± 2 −9 ± 2 −9 ± 2 V8i BiphasicMAL (2) 8 weeks +6 ± 2 +9 ± 2 0 ± 5 +9 ± 7 +16 ± 7* +8 ± 3 +6 ± 2 +3 ± 2 +3 ± 2 +2 ± 2 V8i UnchangedKNI (1) 8 weeks −3 ± 2 +3 ± 2 −3 ± 2 −1 ± 4 −4 ± 1 −5 ± 1 −7 ± 2 −10 ± 1* −9 ± 2* −11 ± 2* V8i Unchanged

1Minute ventilation (V8i) each 30 s during 5 min exposure to 15% O2. Responses of three infants were not analyzed due to periodic breathing (WHE and TUR day 8 and 9; TUR 8 weeks).Not all infants were studied at all ages (HIL, MAL, KNI). Ventilatory response is defined in text.2Tests terminated at 4 min.*Significant differences from control are indicated (ANOVA + Student-Newman-Keuls multiple range tests;P < 0.05).

Response to 15% O 2

The most frequent VRMH recorded within the young-est age group (2.5 ± 0.3 days; 4/5 infants) was a sustaineddecrease inV8i (Table 2); one infant of this age (WHE)exhibited a biphasic response. At a mean age of 7.5 ± 0.7days, hypoxia no longer elicited a sustained decrease inV8i: one infant (RED) exhibited a biphasic response, twoinfants showed no response (MCD, HIL), and two infants(TUR, WHE) developed periodic breathing. Periodicbreathing, when observed, occurred following an aug-

mented breath, or ‘‘sigh’’ and persisted at least until thetest was terminated and usually for some time afterwards.At age 8 weeks, three infants exhibited a biphasicVRMH, two infants had no response, and one infant hada persisting periodic response (Table 2).

A comparison of mean responses recorded for theyoungest and oldest infants studied illustrates the devel-opmental change recorded in the VRMH over the first 2months after birth (Fig. 1). The difference between thesustained hypopnea recorded at 2.5 days and the apparentbiphasic response at 8 weeks is evident from theV8i, aswell as theVT/TI responses (Fig. 1). These differenceswere statistically significant (ANOVA forDV8i: ageP <0.001, timeP < 0.001, interactionP 4 0.05; forDVT/TI:ageP < 0.001, timeP < 0.03, interactionP < 0.005).

Fig. 1. Response to 15% O 2 at two ages. Mean ventilatory re-sponse age 2.5 ± 0.3 days ( d) and 8 weeks ( h). Each curve is themean response of five infants studied during quiet sleep. Notethe emergence of a significant VT response by age 8 weeks.Values significantly different from control are indicated by anasterisk (ANOVA).

Fig. 2. Reproducibility of the response to 15% O 2. Duplicateventilatory responses to 15% O 2 elicited from a 3-day-old infant(TUR) during quiet sleep; there was a 60 min interval betweentests. Minute ventilation ( DV*i, upper panel), tidal volume ( DVT,middle), and breathing rate ( Df, lower panel) are shown as per-cent change from control. Each point is the mean (± SEM) of allbreaths during the preceding 30 s. Values significantly differentfrom control are indicated by an asterisk (ANOVA).

Response of the Newborn to Hypoxia 167

Strictly speaking, only theVT/TI response was unequivo-cally biphasic, since the initial increase inV8i was smalland not statistically significant. No mean response wascalculated for infants aged 7 days because of the smalldata set that remained after the exclusion of periodicresponses.

The sustained decrease inV8i recorded from theyoungest infants studied was due to a reducedVT duringsome tests, and a reducedf during other tests (Fig. 2).The initial hyperpnea that characterized the biphasic re-sponse was due principally to an increasedVT. This in-crease inVT was largely maintained during hypoxia; thesubsequent ventilatory roll-off was principally due to adecreasedf.

Reproducibility of Responses

Duplicate VRMHs were elicited during ten studies.Only once were two tests administered during the sameepoch of quiet sleep, separated by a 10 min interval. Ona further nine occasions, the test was repeated duringquiet sleep after an intervening period of either rapid-eye-movement sleep or wakefulness, the mean intervalbetween these tests being 62 ± 20 min (range, 10–213min). On five of these ten occasions, the VRMH was thesame when the test was repeated. On two occasions, theinitial response was biphasic, but a different responsewas observed when the test was repeated (WHE day 3,V8i decreased; HIL day 6,V8i unchanged). On three oc-casions, sustained hypopnea was observed initially, butV8i did not change significantly during the second test(HIL, TUR, and MCD studied on days 2–3). On all theseoccasions, the VRMH assigned (Table 2) was the re-sponse resulting after data from both tests were com-bined.

SaO2Measurements

SaO2tended to fall slowly and progressively over 5 min

for all except 8-week-old infants, for whom the decreasein SaO2

was largely complete by the end of the secondminute of the test (Table 3). Although theSaO2

profilesduring hypoxia achieved within each study were repro-ducible between tests, there were significant differencesbetween the minimumSaO2

levels achieved with the sameFI,O2

at different ages (Table 3; compare RED at age 1.5days and 8 weeks). There was a significant differencebetween meanSaO2

profiles for infants aged 2 days and 8weeks:SaO2

after breathing 15% O2 for 3 min was 92 ±0.9 vs. 95 ± 0.3, and after 5 min it was 90 ± 1.0 vs. 94 ±0.2; theseSaO2

profiles were significantly different bytwo-way ANOVA (by ageP < 0.001; timeP < 0.001;interactionP 4 0.5).

PetCO2Changes During Hypoxia

Sustained hypopnea during hypoxia was associatedwith a significant increase inPetCO2

at 5 min for allsubjects (Table 4). The biphasic VRMH, however, wasbroadly characterized by reciprocal changes inV8i andPetCO2

: during the initial hyperpnea thePetCO2decreased

significantly, while during the subsequent decrease inV8i, there was an increase inPetCO2

levels, although thisdid not always return to the pre-hypoxic control level bythe end of the fifth minute of the test (Table 4).

DISCUSSION

It has been claimed that the transition from an imma-ture (biphasic) to a mature (sustained hyperpnea) venti-latory response to sustained hypoxia is well advanced bythe end of the second postnatal week for infants born atterm4,7 and slightly later for premature infants.8 A sub-stantially revised view of the postnatal time course ofdevelopment of this response is indicated by our find-ings. This is based partly on a confirmation of old find-ings—that hypoxia results in a sustained fall in ventila-tion under certain conditions15,16—as well as a previ-ously unreported and rather novel finding that arelatively immature (biphasic) response can still be elic-ited from infants 2 months after birth. These observationssuggest that the response of the human infant to hypoxiamatures more slowly than previously thought4,7 and at aslower pace than is believed to occur for other speciesstudied at comparable postnatal ages.17

Environmental conditions have historically been care-fully controlled during studies of the response to hypox-ia, in view of the recognized effect of temperature onventilation.18 Term infants who are studied naked or par-tially clothed soon after birth reportedly exhibit a bipha-sic VRMH in a warm, or thermoneutral environmental(31°–34°C) but a sustained hypopnea at environmentaltemperatures below 28°C.7,15,16 In practice, healthy in-fants experience an environment that probably lies some-where between these two extremes—air (or room) tem-perature may be significantly below thermoneutrality,but clothes and bedding provide a layer of insulation.Evidence indicates that despite the ‘‘cool’’ surroundings,infants nursed in this fashion may actually be close tothermal balance.19 Our findings indicate that for newborninfants studied under these conditions, mild hypoxiaellicts a sustained fall in ventilation, as suggested previ-ously.7,15,16Ceruti7 found that this response was mature(hyperpnea was sustained) at the end of the first postnatalweek, regardless of whether infants were studied in awarm or cool environment. Although we could not con-firm this, there was nevertheless evidence that a substan-tial change in response had occurred by this age, sincethe hypopnea elicited on days 2–3 was absent when the

168 Cohen et al.

TABLE 3— SaO2During Hypoxia1

Subject(age at study)

SaO2during inhalation of 15% O2 (Torr; mean ± sem)

Control 309 609 909 1209 1509 1809 2109 2409 2709 3009

MCD (3 days) 100 ± 0.1 100 ± 0.1 99 ± 0.2* 98 ± 0.2* 97 ± 0.3* 97 ± 0.2* 96 ± 0.3* 96 ± 0.3* 96 ± 0.3* 95 ± 0.4* 94 ± 0.5*TUR (3 days) 98 ± 0.3 98 ± 0.2 96 ± 0.2* 95 ± 0.2* 95 ± 0.2* 94 ± 0.2* 92 ± 0.2* 92 ± 0.2* 91 ± 0.5* 91 ± 0.7* 90 ± 0.2*HIL (2 days) 96 ± 0.3 97 ± 0.1* 96 ± 0.2 94 ± 0.2* 92 ± 0.2* 92 ± 0.2* 92 ± 0.1* 92 ± 0.3* 93 ± 0.2* 91 ± 0.2* 91 ± 0.3*RED (1.5 days) 97 ± 0.2 95 ± 0.3* 93 ± 0.3* 92 ± 0.3* 90 ± 0.5* 89 ± 0.3* 87 ± 0.2* 86 ± 0.2* 85 ± 0.2* 85 ± 0.2* 84 ± 0.3*WHE (3 days) 99 ± 0.1 99 ± 0.1 98 ± 0.2* 97 ± 0.2* 96 ± 0.1* 94 ± 0.2* 95 ± 0.2* 94 ± 0.1* 94 ± 0.2* 93 ± 0.2* 93 ± 0.2*MCD (7 days) 99 ± 0.2 97 ± 0.3* 94 ± 0.3* 95 ± 0.3* 93 ± 0.3* 93 ± 0.4* 93 ± 0.3* 95 ± 0.7* 94 ± 0.3* 93 ± 0.2* 90 ± 0.4*HIL (6 days) 99 ± 0.1 95 ± 0.3* 94 ± 0.5* 96 ± 0.3* 94 ± 0.3* 94 ± 0.3* 94 ± 0.2* 93 ± 0.3* 93 ± 0.2* 92 ± 0.3* 92 ± 0.4*RED (8 days) 98 ± 0.1 96 ± 0.6* 98 ± 0.2* 97 ± 0.3* 97 ± 0.4* 97 ± 0.5* 96 ± 0.2* 96 ± 0.1* 96 ± 0.1* 95 ± 0.3* 96 ± 0.2*MCD (8 weeks) 98 ± 0.1 97 ± 0.2* 96 ± 0.2* 96 ± 0.1* 96 ± 0.1* 95 ± 0.2* 94 ± 0.3* 94 ± 0.1* 94 ± 0.1* — —RED (8 weeks) 100 ± 0.1 97 ± 0.2* 97 ± 0.1* 96 ± 0.4* 97 ± 0.3* 96 ± 0.3* 96 ± 0.3* 95 ± 0.4* 96 ± 0.1* 95 ± 0.3* 95 ± 0.3*WHE (8 weeks) 100 ± 0.1 98 ± 0.2* 97 ± 0.1* 96 ± 0.1* 97 ± 0.6* 97 ± 0.5* 96 ± 0.2* 96 ± 0.1* 96 ± 0.2* 96 ± 0.1* 95 ± 0.1*MAL (8 weeks) 99 ± 0.1 97 ± 0.2* 95 ± 0.1* 95 ± 0.3* 95 ± 0.4* 94 ± 0.6* 94 ± 0.3* 94 ± 0.1* 93 ± 0.2* 93 ± 0.2* 94 ± 0.1*KNI (8 weeks) 99 ± 0.1 96 ± 0.3* 95 ± 0.1* 94 ± 0.2* 94 ± 0.2* 93 ± 0.2* 94 ± 0.2* 93 ± 0.2* 93 ± 0.1* 93 ± 0.1* 93 ± 0.1*

1Arterial O2 saturation (SaO2, Nellcor N-100 pulse oximeter) during a 60 s control period, and during successive 30 s epochs of a 5 min period breathing 15% O2. Details as for Table 1.

TABLE 4— PetCO2During Inhalation of 15% O2

1

Subject(age at study)

PetCO2during inhalation of 15% O2 (Torr; mean ± sem)

Control 309 609 909 1209 1509 1809 2109 2409 2709 3009

MCD (3 days) 35.9 ± 0.2 36.9 ± 0.1* 37.3 ± 0.2* 38.0 ± 0.2* 37.8 ± 0.2* 38.5 ± 0.1* 39.1 ± 0.2* 39.5 ± 0.1* 39.7 ± 0.1* 39.0 ± 0.3* 39.7 ± 0.1*TUR (3 days) 41.6 ± 0.1 41.5 ± 0.2 41.3 ± 0.2 41.9 ± 0.2 42.0 ± 0.1 42.6 ± 0.2* 43.1 ± 0.2* 43.5 ± 0.1* 43.5 ± 0.1* 43.6 ± 0.1* 44.0 ± 0.1*HIL (2 days) 30.5 ± 0.3 30.8 ± 0.2 30.3 ± 0.3 29.1 ± 0.5* 30.2 ± 0.2 30.5 ± 0.3 30.5 ± 0.5 29.4 ± 0.5 31.2 ± 0.4 31.0 ± 0.7 32.3 ± 0.1RED (1.5 days) 34.8 ± 0.1 35.4 ± 0.3 35.7 ± 0.2* 35.6 ± 0.2 35.8 ± 0.2* 35.5 ± 0.2 35.5 ± 0.2 35.2 ± 0.3 35.9 ± 0.3* 35.6 ± 0.2 36.2 ± 0.3*WHE (3 days) 33.8 ± 0.1 33.0 ± 0.1* 32.8 ± 0.1* 32.9 ± 0.1* 33.3 ± 0.1* 33.6 ± 0.1 34.2 ± 0.2* 34.4 ± 0.1* 34.6 ± 0.1* 34.9 ± 0.1* 35.0 ± 0.1*MCD (7 days) 42.3 ± 0.1 42.0 ± 0.2 42.0 ± 0.2 41.2 ± 0.2* 41.6 ± 0.2 41.8 ± 0.2 42.3 ± 0.3 43.3 ± 0.3* 42.8 ± 0.3 41.9 ± 0.5 42.2 ± 0.5HIL (6 days) 41.4 ± 0.1 40.2 ± 0.6* 39.2 ± 0.5* 41.1 ± 0.1 41.0 ± 0.1 40.9 ± 0.2 40.9 ± 0.1 41.2 ± 0.2 40.8 ± 0.3 41.3 ± 0.3 40.8 ± 0.3RED (8 days) 37.3 ± 0.1 37.0 ± 0.2 37.0 ± 0.2 36.3 ± 0.1* 36.9 ± 0.1 36.6 ± 0.3* 37.5 ± 0.1 37.6 ± 0.2 37.8 ± 0.1 38.1 ± 0.2* 37.5 ± 0.2MCD (8 weeks) 42.2 ± 0.1 41.9 ± 0.1* 41.4 ± 0.1* 41.1 ± 0.1* 41.0 ± 0.1* 41.3 ± 0.1* 41.9 ± 0.1* 41.7 ± 0.1* 41.6 ± 0.1* — —RED (8 weeks) 38.4 ± 0.1 37.1 ± 0.2* 36.4 ± 0.2* 36.6 ± 0.3* 36.4 ± 0.1* 36.6 ± 0.2* 36.6 ± 0.3* 37.1 ± 0.1* 36.9 ± 0.8* 37.3 ± 0.2* 37.6 ± 0.2*WHE (8 weeks) 35.1 ± 0.1 34.6 ± 0.1* 34.7 ± 0.1* 34.6 ± 0.1* 34.6 ± 0.1 34.2 ± 0.1* 34.0 ± 0.1* 34.0 ± 0.1* 34.0 ± 0.1* 34.1 ± 0.1* 34.1 ± 0.1*MAL (8 weeks) 40.4 ± 0.1 40.6 ± 0.2 40.3 ± 0.1 39.3 ± 0.2* 38.8 ± 0.3* 39.3 ± 0.3* 39.3 ± 0.1* 39.7 ± 0.1* 39.9 ± 0.1* 40.1 ± 0.1 39.8 ± 0.1KNI (8 weeks) 35.9 ± 0.3 36.6 ± 0.2 35.9 ± 0.2 35.8 ± 0.3 35.9 ± 0.3 36.0 ± 0.2 36.4 ± 0.2 36.4 ± 0.4 37.3 ± 0.2* 37.1 ± 0.2* 36.6 ± 0.3

1End-tidalPCO2(PetCO2

) levels during a 60 s control period, and during successive 30 s epochs of a 5 minute period breathing 15% O2. Details as for Table 1.

same infants were re-tested approximately 5 days later.This is broadly consistent with the view that the periph-eral chemoreflex of the newborn is refractory to hypoxiaimmediately after birth and gradually becomes more sen-sitive with increasing postnatal age.20,21 Recently it wassuggested that this resetting process may be more or lesscomplete within 48 h of birth, based on the similarity ofthe ventilatory responses of 2-day and 7-week-old infantsto an oscillatingPaO2

stimulus.22 We did find a signifi-cant difference between mean ventilatory responses atthese ages, as well as a relatively immature response at 8weeks. This suggests that postnatal re-ordering of thecomponents that determine the overall response to hyp-oxia may be quite slow and is certainly not completewithin 48 h, perhaps not for several more months afterbirth. The different time courses of maturation suggestedby these two studies may reflect the specific character-istics of the stimuli presented.

Some of the differences between our findings andthose reported may be attributed to the environmentalconditions at the time of the studies or the test protocols(above). Where there were similarities in protocol, how-ever, it may be important to examine the manner inwhich tests were administered, as well as the methodsused to analyze the data. We analyzed substantially moredata than either of the two previous serial studies of theterm infant’s response to 15% O2 (26 vs. 4–6 tests). Theduration of hypoxia was also slightly greater during ourstudies (5 vs. 3 min) which we felt permitted better over-all resolution of the ventilatory changes elicited. A par-ticular difficulty we encountered was the absence fromprevious reports of a thorough statistical analysis of theventilatory changes described. We avoided qualitativedescriptions of the VRMH, instead relying on objective,statistical criteria to define each response. These criteriawere not particularly innovative but merely codified thespectrum of responses recognized and reported by earlierinvestigators. There is no reason to believe that the in-fants we studied were unusual—they were all healthy,from uncomplicated pregnancies. Moreover, theV8i, VT,f, and PetCO2

recorded at rest were within the normalranges reported for infants of this age.23 It is important torecognize, however, that changes in the response to hyp-oxia were apparent within a reasonably rapid time frame,between days 2–3 and 7–9. This change may not havebeen discerned if all data from studies undertaken be-tween days 2–9 had been pooled. By deriving responsesfrom too heterogeneous an age group, previous investi-gators may not have provided an accurate picture of theresponse at a discrete point in time.2

It is believed that input from the chemoreceptors iscritical to drive breathing during quiet sleep.9 Studies ofpreterm infants indicate that a biphasic VRMH occursonly during wakefulness or REM sleep, and it has beensuggested that the emergence of a mature, sustained hy-

perpneic response is linked to the postnatal increases inthe proportion of quiet sleep.24 To date term infants havebeen studied only during wakefulness, or during sleepwhen the state was not specifically defined. Our studiesindicate that term infants do have a prominent biphasicresponse during quiet sleep. Changes in the componentsof V8i during this state varied somewhat from those re-ported for awake or ‘‘sleeping’’ infants. Although sus-tained hypopnea is reportedly due almost entirely due tochanges inVT,15 we found that with some exceptions(Fig. 2), it was principally due to a decrease inf (Fig. 1).The biphasic response is reportedly due to failure to sus-tain the initialVT response.2,15,16We found that althoughthe initialVT increase was reasonably well sustained dur-ing hypoxia, V8i decreased because the breathing rateslowed significantly (Fig. 1). Preterm infants studied dur-ing quiet sleep behaved similarly.24,25

Ventilatory responses to CO2 tend to be relatively re-producible during QS.26 Test-to-test variability of theVRMH appears to be greater, since in this study less thanhalf the duplicate responses we elicited agreed. Althoughthis may have been exaggerated by our use of strict limitsto define responses, QS is unlikely to be a homogeneousstate,27 and the variability of the response may reflect theprotracted duration of the stimulus. Periodic breathingpresented particular difficulties in the calculation of theVRMH, and for this reason we chose to consider it as avalid but discrete entity when tabulating responses. Com-monly observed in newborn infants,28 periodic breathingmay occur (as we found) following a sigh29,30 and withincreased incidence during experimental hypoxia.3,31

The mechanism of the fall inV8i during experimentalhypoxia is the subject of ongoing controversy.5 There hasbeen speculation that in the newborn, as in the fetus,metabolic rate falls during hypoxia, causing tissuePCO2

levels to fall, reducing one of the tonic components ofrespiratory drive and precipitating a decrease inV8i.17,32

A sustained and significant fall inPACO2during hypoxia

has in fact been reported.4,7,24 Interestingly, preventingany fall in tissuePCO2

by adding CO2 to a hypoxic in-spirate does not abolish the biphasic response.16 Our datasuggest that the relationship betweenPetCO2

andV8i wasmore often reciprocal:PetCO2

tended to increase whenV8idecreased, and vice-versa. This reciprocal inter-relationship suggests that there was little change in meta-bolic rate during the 5 min period of hypoxia that ourinfants experienced. Metabolic studies are yet to arrive ata consensus about whether or not significant metabolicchanges occur during sustained hypoxia at this youngage.32,33 This may depend on the severity of hypoxia; ithas been suggested that breathing 15% O2 may not besufficient to provoke a hypometabolic response.32

In a true steady-state test, there should be a rapid(within 1–2 min) initial reduction inSaO2

(or PaO2, if that

is measured) to a level that is maintained for the duration

170 Cohen et al.

of the test.6 It has long been assumed that this is thegeneral profile of the changes experienced by infantsfollowing a step reduction inFI,O2

; in fact, our data in-dicate that this depends on the age at the time of thestudy. At age 2 months, breathing 15% O2 did result in ashallow, stepwise, ‘‘steady-state’’ fall inSaO2

, but whenthe same test was administered to infants aged 2–3 days,there was more likely to be aprogressive(and eventuallymore profound) fall inSaO2

. Although this may in part beexplained by hemodynamic factors, it must also reflectthe ‘‘closed loop’’ configuration of the system (i.e., theinterdependence of changes inSaO2

andV8i). Under suchcircumstances, a fall inV8i should accentuate the fall inSaO2, which is what was observed; hypoxemia wasgreatest at 2–3 days, when a progressive hypopnea de-veloped while the infant was breathing 15% O2; this wasless at 2 months of age, when infants responded with atransient hyperpnea.

We conclude that the classical, steady-state test ofhypoxic responsiveness provides too variable a stimulusto be of great value in evaluating postnatal changes in O2

chemoreflex. Not only did the extent of the overall fall inSaO2

vary [at the end of 5 min, the meanSaO2varied by as

much as 11%, depending on the age of the infant studied(Table 3)], but the profile of changes also varied: some-times theSaO2

fell in a shallow, stepwise fashion, whileon other occasions slow, progressive desaturation wasevident. Such variable profiles may themselves have asignificant impact on the response elicited.34,35 Underthese circumstances, it is neither realistic nor reliable tocompare responses elicited from infants of different post-natal ages. A more reproducible stimulus (and more use-ful comparative information) might be obtained if theFI,O2

is adjusted independently to achieve a predeter-mined, ‘‘target’’ SaO2

that does not vary with the venti-latory response.6

REFERENCES

1. Hanson M, Calder N, Watanabe T, Kumar P. The drive to breatheand the sudden infant death syndrome. Acta Paediatr Suppl. 1993;389:47–49.

2. Cross KW, Warner P. The effect of inhalation of high and lowoxygen concentrations on the respiration of the newborn infant. JPhysiol (Lond). 1951; 114:283–295.

3. Cross KW, Oppe´ TE. The effect of inhalation of high and lowconcentrations of oxygen on the respiration of the premature in-fant. J Physiol (Lond). 1952; 117:38–55.

4. Brady JP, Ceruti E. Chemoreceptor reflexes in the newborn infant:Effects of varying degrees of hypoxia on heart rate and ventilationin a warm environment. J Physiol (Lond). 1966; 184:631–645.

5. Johnston BM, Walker DW. Respiratory responses to changes inoxygen and carbon dioxide in the perinatal period. In: JohnstonBM, Gluckman PD, eds. Respiratory Control and Lung Develop-ment in the Fetus and Newborn. Ithaca: Perinatology Press, 1986:279–310.

6. Easton PA, Slykerman LJ, Anthonisen NR. Ventilatory response

to sustained hypoxia in normal adults. J Physiol (Lond). 1986;61:906–911.

7. Ceruti E. Chemoreceptor reflexes in the newborn infant: Effect ofcooling on the response to hypoxia. Pediatrics. 1966; 37:556–564.

8. Rigatto H, Brady JP, de la Torre Verduzco R. Chemoreceptorreflexes in preterm infants: I. The effect of gestational and post-natal age on the ventilatory response to inhalation of 100% and15% oxygen. Pediatrics. 1975; 55:604–613.

9. Sullivan CE. Breathing in sleep. In: Orem J, Barnes D, eds. Physi-ology in Sleep. New York: Academic Press, 1980:213–271.

10. Johnson P, Andrews DC. Thermometabolism and cardiorespira-tory control during the perinatal period. In: Beckerman RC,Brouillette RT, Hunt CE, eds. Respiratory Control Disorders inInfants and Children. Baltimore, Williams & Wilkins, 1992:76–88.

11. Cohen G, Henderson-Smart DJ. A modified rebreathing method tostudy the ventilatory response of the newborn to carbon dioxide.J Dev Physiol. 1990;14:295–301.

12. Prechtl HFR. The behavioural states of the newborn infant (areview). Brain Res. 1974; 76:185–212.

13. Anders T, Emde R, Parmelee A. A Manual of Standardised Ter-minology, Techniques and Criteria for Scoring of States of Sleepand Wakefulness in Newborn Infants. Los Angeles: UCLA BrainInformation Service, 1971.

14. Barrington KJ, Finer NN, Wilkinson MH. Progressive shorteningof the periodic breathing cycle duration in normal infants. PediatrRes. 1987; 21:247–251.

15. Miller HC, Smull NW. Further studies on the effects of hypoxiaon the respiration of newborn infants. Pediatrics. 1955; 16:93–102.

16. Brady JP, Dunn PM. Chemoreceptor reflexes in the newborn in-fant: Effects of CO2 on the ventilatory response to hypoxia. Pe-diatrics. 1970;45:206–215.

17. Mortola J, Rezzonico R, Lanthier C. Ventilation and oxygen con-sumption during acute hypoxia in newborn mammals: A compara-tive analysis. Respir Physiol. 1989; 78:31–43.

18. Hill JR. The oxygen consumption of newborn and adult mammals.Its dependence on the oxygen tension in the inspired air and on theenvironmental temperature. J Physiol (Lond). 1959; 149:346–373.

19. Hey EN, O’Connell B. Oxygen consumption and heat balance inthe cot-nursed baby. Arch Dis Child. 1970; 45:335–343.

20. Kumar P, Hanson MA. Re-setting of the hypoxic sensitivity ofaortic chemoreceptors in the newborn lamb. J Dev Physiol. 1989;11:199–206.

21. Marchal F, Bairam A, Haouzi P, Crance JP, Di Giulio C, Vert P,Lahiri S. Carotid chemoreceptor response to natural stimuli in thenewborn kitten. Respir Physiol. 1992; 87:183–193.

22. Calder NA, Williams BA, Kumar P, Hanson MA. The respiratoryresponse of healthy term infants to breath-by-breath alterations ininspired oxygen at two postnatal ages. Pediatr Res. 1994; 35:321–324.

23. Hathorn MKS. The rate and depth of breathing in new-born in-fants in different sleep states. J Physiol (Lond). 1974; 243:101–113.

24. Rigatto H, Kalapesi Z, Leahy FN, Durand M, MacCallum M,Cates D. Ventilatory response to 100% O2 and 15% O2 duringwakefulness and sleep in preterm infants. Early Hum Dev. 1982;7:1–10.

25. Alvaro R, Alvarez J, Kwiatkowski K, Cates D, Rigatto H. Smallpreterm infants (<1500g) have only a sustained decrease in ven-tilation in response to hypoxia. Pediatr Res. 1992; 32:403–406.

26. Cohen G, Xu C, Henderson-Smart DJ. Ventilatory response of thesleeping newborn to CO2 during normoxic rebreathing. J ApplPhysiol. 1991; 71:168–174.

27. Shulze K, Kairam R, Stefanski M, Sciacca R, Bateman D, Dell R,

Response of the Newborn to Hypoxia 171

James LS. Spontaneous variability in minute ventilation, oxygenconsumption and heart rate of low birth weight infants. PediatrRes. 1981; 15:1111–1116.

28. Glotzbach SF, Ariagno RL. Periodic breathing. In: BeckermanRC, Brouillette RT, Hunt CE, eds. Respiratory Control Disordersin Infants and Children. Baltimore: Williams & Wilkins, 1992:142–160.

29. Carse E, Wilkinson AR, Whyte PL, Henderson-Smart DJ, JohnsonP. Oxygen and carbon dioxide tensions, breathing and heart rate innormal infants during the first six months of life. J Dev Physiol.1981; 3:85–100.

30. Fleming P, Goncalves AL, Levine MR, Woollard S. The devel-opment of stability of respiration in human infants: Changes inventilatory responses to spontaneous sighs. J Physiol (Lond).1984; 347:1–16.

31. Rigatto H, Brady JP. Periodic breathing and apnea in preterminfants. I. Evidence for hypoventilation possibly due to centralrespiratory depression. Pediatrics. 1972; 50:202–218.

32. Oliver TK, Karlberg P. Gaseous metabolism in newly born humaninfants. Am J Dis Child. 1963; 105:427–435.

33. Cross KW, Flynn DM, Hill JR. Oxygen consumption in normalnewborn infants during moderate hypoxia in warm and cool en-vironments. Pediatrics. 1966; 37:565–576.

34. Mahutte CK, Rebuck AS. Influence of rate of induction of hyp-oxia on the ventilatory response. J Physiol (Lond). 1978; 284:219–227.

35. Bureau MA, Coˆte A, Blanchard PW, Hobbs S, Foulon P, Dalle D.Exponential and diphasic ventilatory response to hypoxia in con-scious lambs. J Appl Physiol. 1986; 61:836–842.

172 Cohen et al.