Ventilatory response of the sleeping newborn to CO

0 downloads 0 Views 2MB Size Report
J. Appl. Physiol. 58: 1982-1987, 1985. FINER, N. N., I. F. ABROMS, AND H. W. TAEUSCH. Ventilation and sleep state in newborn infants. J. Pediatr. 89: 100-108 ...
Ventilatory response of the sleeping newborn to CO, during normoxic rebreathing GARY COHEN, CATHY XU, AND DAVID HENDERSON-SMART Department of Perinatal Medicine, King George V Memorial Hospital, Sydney, Neul South Wules 2050, Australia CUHEN,GARY,CATHYXU,ANDDAVIDHENDERSON-SMART. Ventilatory response of the sleeping newborn to CO, during normoxie rebreathing. J. Appl. Physiol. 71(l): 168-174,1991.-The

ventilatory response of the newborn to CO, was studied using a rebreathing method that minimized changes in arterial PO, during the test. The aim was to study the variability of the ventilatory response to CO, and take this into account to assess the relative magnitude of the response to COZ during rapid-eyemovement (REM) sleep and quiet sleep (QS). Five full-term babies aged 4-6 days were given 5% CO, in air to rebreathe for 1.5-3 min. 0, was added to the rebreathing circuit to maintain arterial 0, saturation and transcutaneous PO, (Ptco,) at prerebreathing levels. Tests were repeated four to five times in REM sleep and QS. Mean Ptc,, levels varied between individuals but were similar during REM sleep and QS tests for each subject. The mean coefficient of variability of the ventilatory response was 35% (range l&77%) during QS and 120% (range 32-220%) during REM sleep. Ptc,, fluctuations during tests [6.0 & 3.0 (SD) Torr, range l-13 Torr] were not correlated with ventilatory response. Overall the ventilatory response was significantly lower in REM sleep than in QS (12.2 t 3.0 vs. 38.7 t 3.0 ml. min-’ 9Torr-l . kg-l, P < 0.001; 2-way analysis of variance) due to a small (nonsignificant) fall in the tidal volume response and a significant fall in breathing rate. In 12 REM sleep tests there was no significant ventilatory response; mean inspiratory flow increased significantly during 8 of these 12 tests. We conclude that there is a significant decrease in the ventilatory response of the newborn to COZ rebreathing during REM sleep compared with QS. sleep state; full-term

infants; chemoreceptors,

reproducibility

been reported. The rebreathing practical

and theoretical

method enjoys several

advantages

over

the steady-

state method (27); however, the results obtained in the newborn are confused by rebreathing normally being carried out under hyperoxic (40% 0,) conditions. Although this has been necessary to prevent the subject from becoming

hypoxemic

in the course of the test, high

inspired concentrations of 0, cannot be regarded as inert, because a biphasic ventilatory response of the newborn to hyperoxia has been demonstrated (7). Ideally, therefore, the ventilatory response to CO, should be measured in air. Recently we reported a modified rebreathing technique to measure the ventilatory response of full-term and premature infants to CO, under approximately normoxie conditions (6). This was achieved using a CO,-air mixture

as the rebreathing

gas, to which

0, was added

during the test at the estimated metabolic consumption rate. The results indicated that, as in adults (2, lo), there is considerable test-to-test variability of the ventilatory response to CO,. Previous reports of the ventilatory responses during different sleep states have not considered this and only one test was conducted in each state or one ventilatory test was performed per subject. In this study, we examined the variability and sleep-state-related differences in the ventilatory response of healthy full-term infants to CO, by performing multiple tests in each sleep state. MATERIALS

AND METHODS

sponse was measured during steady-state CO, breathing (8, 13) or CO, rebreathing (16, 22). Furthermore, repli-

Five healthy full-term infants were selected for study from the general nurseries. Details of the gestational age and weight at birth and postnatal age are given in Table 1. Studies were carried out between midmorning and midafternoon in a laboratory attached to the special care nursery, beginning ~30 min after a scheduled feeding. Infants were clothed and studied as they lay supine in open cots. Room temperature was maintained at 2325°C; skin (axillary) temperature was measured at the beginning and end of the study. The following variables were recorded: the electroencephalogram (EEG; from electrodes placed on the vertex and the mastoid process), electrooculogram (EOG; from electrodes placed at the outer canthus and referred to the mastoid process or from a piezoelectric crystal taped over the eyelid), chest wall movement (from a mercury-in-

cate tests to assess the test-to-test

rubber

OUR UNDERSTANDING maintains

the newborn rudimentary

of the reflex processes by which effective

breathing

efforts

is still

(5). Some evidence suggests that infants

with breathing disturbances during sleep have different reflex responses to hypercapnia and hypoxia than do apparently normal infants (15). Unfortunately, when much

of this evidence was gathered, it was often not appreciated that the level of arousal greatly influences the response to specific stimuli (5,24,31). As a result, there has been a need to reexamine these findings in relation to behavioral state. Recent studies comparing the ventilatory response of the newborn to CO, during rapid-eyemovement (REM) sleep and quiet sleep (QS) reached different

168

conclusions,

depending

on whether variability

0161~7567/91

the rehave not

$1.50 Copyright

strain

0 1991 the American

gauge taped

Physiological

Society

around

the chest at nipple

RESPONSE TABLE

OF THE

NEWBORN

1. Subject data

Subj

Sex

GA, wk

PNA, days

Birth Wt, g

LU SA GR VO TO

M

40 38 39 40 40

4 5

3,710 3,860 3,114 2,115 2,900

F

M F F

GA,

5 6 5

gestationalage;PNA, postnatal age.

level), arterial 0, saturation (Sao,; Nellcor model N-100 pulse oximeter attached to the foot; response time 3 s), heart rate (generated from the pulse of the oximeter by means of a Grass 7 P4 preamplifier), end-tidal Pcoz (PE%oz ; Engstrom Eliza CO,-0, analyzer; sampling rate -80 mUmin), and transcutaneous PO, (Ptc,,; Roche 820 transcutaneous electrode, attached to the abdominal skin). Airflow was measured using a miniaturized mask-pneumotachograph to which a Validyne model MP-45 differential pressure transducer was attached with flexible polyvinyl tubing. The pneumotachograph was calibrated at the start and finish of each study; linear interpolation between initial and final calibration runs was used to minimize error arising from drift (which was assumed to be linear). The calibration procedure and characteristics of the device have been described in detail (6). The rebreathing technique has been described (6). A small detachable fiberglass mask at one end of the pneumotachograph covered the baby’s nose and upper lip. A ring of Silastic putty between the mask and the baby’s skin held the system in place and prevented leaks. The pneumotachograph assembly )was attached lo-20 min after a scheduled feeding, and the baby was then allowed to sleep lying in a supine or slightly lateral position. Briefly, to begin a rebreathing run, the pneumotacograph was switched from air to a bag containing 5% CO, in air. The initial volume of the bag was determined by trial and error to achieve rapid equilibration between CO, in the bag and lungs, without impeding breathing. Gas withdrawn for CO, and 0, analysis was recirculated; before being returned to the bag, 0, was added at ~7 ml min-l kg body wt-l to replace that consumed. 0, inflow was adjusted (if necessary) to maintain Sa,, and Ptco, at prerebreathing levels. Tests were of 1.5 to 3-min duration, repeated four to five times each in REM sleep and QS, after the PET,,, , Ptc,,, and tidal volume were judged to have returned to prerebreathing levels. QS was determined by the presence of a trace-alternant EEG pattern, regular breathing movements, and the absence of eye and gross body movements except for brief startles; REM sleep was determined by the lower-voltage EEG, frequent eye and body movements, and the irregular breathing pattern. Tests were, whenever possible, terminated before arousal occurred. Tests were abandoned or excluded if the Ptc,, was excessively high or low in relation to the prerebreathing level (which may have occurred if the 0, inflow rate was too high or too low) or if there were air leaks or the subject aroused. The procedure outlined here was approved by the Ethics Review Committee of this hospital, and all studies were carried l

l

TO CO, DURING

SLEEP

169

out with the informed consent of the parents and the attending physician. Analysis. All signals were recorded on paper with a 16-channel polygraph (Grass Instruments). The analogue signals of airflow, PET,,,, and Sa, were converted to digital signals by computer and stored for later analysis (IBM-compatible personal computer with an Analog Devices RTI-815 analog-to-digital converter). The first 25-30 s of each test (representing equilibration time) were deleted from the analysis; if there remained ~50 s for analysis, the run was discardec$ Tidal volume (VT), instantaneous minute ventilation (VI), mean inspiratory flow (VT/TI), inspiratory (TI) and expiratory time (TE), and instantaneous breathing rate (f) were calculated for each breath. Plots of PETIT, vs. instantaneous VT, VI, VT/TI, and f were constructed for each test, and the lines of best fit were calculated by linear leastsquares regression. Regression coefficients for VI, VT, and VT/TI vs. PETIT, were normalized for body weight. The Ptc,, was measured at 10-s intervals during the analyzed portion of the run. On the basis of comparisons we made previously during simultaneous measurements of Sao, and Ptc,,, an -8-s time lag was estimated between Sa,, and Ptc,, changes. To allow for this 8-s delay, the first Ptc,, measurement during rebreathing was made 8 s after the start of the analysis and subsequently at 10-s intervals. The prerebreathing or control Ptcoz was measured 10 s before the beginning of each test. Two-way analysis of variance (by subject and sleep state) was used for statistical comparison of responses in REM sleep vs. QS; significance was taken at P < 0.05. RESULTS

The number of tests performed in each epoch of REM sleep or QS varied according to the duration of each epoch. Often it was possible to perform three tests within a single epoch; however, to achieve our objective of four or five runs in each state, two epochs were required on average (range l-3). There were 46 successful tests, 22 in REM sleep and 24 in QS. The number of tests per subject, together with mean changes in PET,~, in the course of testing, are given in Table 2. PETITE changes were similar between REM sleep and QS for the same individual, except subject VO, in whom PET,,, increased more rapidly in REM sleep tests than in QS tests. The inspired 0, concentration often fell to 19-17% during the course of rebreathing tests; however, the developing hyperventilation, combined with the low flow of added 0,, prevented large changes in Sao, or Ptc,, during tests. Mean Sa,, during rebreathing tests was 94 t 2.5% (SD). Mean Ptc,, levels during rebreathing varied from subject to subject, reflecting differences in the measured resting Ptc,, . Mean Ptc,, before and during rebreathing was similar within the same subject (prebreathing Ptc,, was slightly lower during REM sleep for subject SA) (Table 2). The mean interval between tests was 270 t 76 s. Skin temperature remained within the range 36.437.0°C (mean 36.8 t 0.04”C). The response during REM sleep was more variable on a breath-to-breath basis than that recorded during QS (Fig. 1). The slope of the ventilatory response to CO,

170 TABLE

RESPONSE

of rebreathing

2. Details

OF THE

NEWBORN

TO CO, DURING

SLEEP

tests by subject and state -‘Qo,, Tom

Ptc,, Torr cIPET~~, /dt,

Subj

State

Minimum

Maximum

Torrlmin

LU

REM

64.5t1.3 60.9t0.7 58.2t0.6 59.5t0.4 54.8k1.4

REM

53.5kO.8 52.0&0,3 48.8k0.3 49.4zko.4 46.3t0.6 46.3kO.4 49.8t0.4 50.73-0.4 51.6kO.6

QS

51.1t0.5

5.420.3 5.6k0.2 7.2k0.2 7.4kO.3 6.71trO.4 6.2rtrO.l 9.8-1-0.8 7.7t0.3 6.6k0.2 6kkO.l

QS SA

REM

QS GR

REM

QS vo

REM

QS TO

56.1kO.4 60.4kO.9

61.OkO.7

62.0t0.9 61.1k0.7

Duration, Control

87.522.0 90.2,+0.2 60.6t1.3 67.4t0.6 89.5k2.1 85.0-+1.0 72.8k1.3

69.61b1.5 65.3k1.4 66.0t1.5

Rebreathing

S

n

88.6t1.4 87.Ok1.7

lE.Ok6.0 89.0t8.0 77.0t6.0 82.0t4.0 74.0+11.0 85.0t4.0 67.0t5.0

4 5 5 5 4 5 5 5 4 4

68.6-10.3 71.0k1.2 90.720.9 85.222.4 70.7k1.9 69.5zkl.l

Values are means 2 SE; n, no. of tests. REM and QS, rapid-eye-movement and quiet sleep. dPET&dt, transcutaneous PO,; Duration, duration of analyzed portion of test (excludes equilibration period).

varied from test to test within the same subject. Responses during QS were generally greater in magnitude and less variable than responses measured during REM sleep (Fig. 2). Mean ventilatory responses calculated for each subject were consistently lower in REM sleep (Fig. 3). During QS, VT always increased significantly; f either did not change (14 tests) or increased (10 tests). During REM sleep, VT increased during all except one test; f increased (1 test), did not change (12 tests), or decreased (9 tests). Variability, expressed as the coefficient of variation of the mean response for each subject, was considerably greater in REM sleep than QS (Table 3). VI did not change (i.e., regression coefficients of PETITE vs. 01 were not significantly different from zero) during 12 of the 22 REM sleep tests; VT/TX did increase significantly during 8 of these 12 tests (Fig. 4). Mean ventilatory changes of the five subjects as a

81.0-r-3.0 91.0t8.0 97Akk2.0

64.2t1.2 66.4kO.9

rate of change of aTCo2;

group during REM sleep and QS are summarized in Table 4. The ventilatory response during REM sleep was 68% less than the QS response. This was due to a small but nonsignificant (18%) decrease in the VT response and a marked tendency for f to decrease to below control levels. Mean ventilatory response lines for REM sleep and QS are plotted for comparison in Fig. 5. The difference between the highest and lowest Ptc,, levels measured during rebreathing tests varied from 1 to 13 Torr (mean 6.0 t 3.0). There was no apparent correlation between the slope or variability of the ventilatory response from test to test and the mean of the corresponding signed fluctuations in Ptc,, from the prerebreathing level. Similarly, there was no relationship between the ventilatory response and differences in the rate of increase of PETIT, from test to test or the duration of the interval between tests.

r

25

01

1

45

50 End-Tidd

1 55 CO2 hmtigl

I 60

d---w5

J 65

50

End-Tidal

2.5

60

I

01 45

50 End-Tfdrl

I

I

60 55 CO2 LmmHg)

I

65

I 50

55 60 CO2 ImmHgl

End-Tidal

I I 60 55 CO, lmmHgt

r

50I

55I End-Tidnl

COz lmmH$

601

654

50I

55I End-Tidal

CO, LmmHg)

FIG. 1. Breath-by-breath plots of end-tidal Pcoz (PET& vs. minute ventilation (VI, left), tidal volume (VT, middle), and breathing rate (f, right) for quiet sleep (QS) test (top) and rapid-eye-movement (REM) sleep test (bottom) for subject TO. Note increased breath-to-breath variability during REM sleep and significant increase in f during QS (T = 0.61,-P c 0.001) but absence of change in f during REM sleep (r = 0.08, P > 0.5). Regression coefficients for PETITE vs. (VI) are similar for both tests (18.6 vs. 17.8 ml min-’ Torr-l . kg-‘) because of increased VT response in REM sleep (0.3 vs. 0.5 ml 9Torr-’ kg-‘, QS vs. REM sleep). l

l

l

Ptco,;

601

I 65

RESPONSE

8 8

OF THE

NEWBORN

0

-20 LU

2.5 r

SA

h

LU

GR

VO

TO

vo

TO

0

SA

GR

2. Minute ventilatory responses ( dk/dPETcoz) and mean inspiratory flow responses [d(VT/TI)/dPETcos] for all 46 rebreathing tests in REM sleep (a) and QS (0). Tests are grouped by subject (abbreviations along abscissa). Note change in scale (ordinate) in bottom panel. FIG.

DISCUSSION

We have found a consistent state-related difference in the ventilatory response of newborn infants to inhaled CO, even when the considerable test-to-test variability is taken into account by repeating tests and comparing means. Recent studies of full-term and preterm infants by use of a hyperoxic rebreathing technique support the conclusion that the ventilatory response to CO, is significantly reduced in REM sleep, compared with the response during QS (16, 22). There are at least two other studies of the newborn however, that disagree. Using a steady-state technique, Davi et al. (8) and Haddad et al. (13) found comparable responses in REM sleep and QS in full-term and premature babies. The steady-state response to CO, is normally calculated as the difference in ventilation between two steady states, established while subjects breathe air and 2 or 3% CO, in air, respectively. A recent study of the time course of ventilatory changes to inhalation of 4% CO, has shown that a new steady ventilatory state will not always be reached within the usual duration of steady-state tests (4-5 min) (20). This raises doubts about the suitability and accuracy of this method, as commonly used. The rebreathing method of measuring CO, sensitivity, by comparison, seems justified in the assumptions it makes and the conclusions it reaches (27,28). On the basis of this method, there would appear to be an emerging consensus that the ventilatory

TO CO2 DURING

171

SLEEP

sensitivity of the human newborn is reduced during REM sleep. In this respect, the newborn resembles the adult (2, 10) and certain species such as the dog (25). Although our findings and those of Honma et al. (16) and Moriette et al. (22) agree about the direction of the change in CO, response during sleep, there are some quantitative differences in calculated responses. Mean values for the ventilatory responses presented in Table 4 are considerably less than those reported by Moriette et al. (22) and are also less than those that can be calculated from the data on full-term babies provided by Honma et al. (16). Similarly, we recorded a larger reduction in response during REM sleep than did either of the two quoted studies (68% vs. 34 and 45%, respectively). This may be partly explained by differences in arterial oxygenation. We did not raise the 0, concentration of the inspired gas above 21% (usually it fell slightly). Any changes in arterial PO, would probably have been minor, if the Ptc,, measurements reliably estimated arterial PO, (19). By comparison, Honma et al. (16) and Moriette et al. (22) conducted rebreathing studies under hyperoxic conditions (40% 0,), during which there would have been large changes in arterial PO,. Ventilatory responses measured under these conditions are likely to reflect a combined effect of raised PO, and PCO, on ventilation (3, 7). During the 1st min of hyperoxic rebreathing, the raised 0, levels would be expected to lead to a relative reduction in peripheral chemoreceptor drive. Moreover, Moriette et al. (22) and Honma et al. (16) did not indicate the variability of the CO, responses of their subjects (Honma et al. made duplicate measurements for 1 subject). If the range of responses we documented for each subject accurately reflects within-subject variability of the ventilatory response to CO, (and this is not certainsee below), it is possible by relying on the results of single tests to either exaggerate or underestimate state-related (or, for that matter, intersubject) differences. The ventilatory response to CO,, as well as being reduced during REM sleep, was more irregular and more variable than was the QS response. This was shown by the greater breath-to-breath variability in VI, VT, f, and TABLE

of ventiihtory

3. Variability

response to CO, cv, %

Subject

State

MVR

MIF

LU

REM

115 29

70 56

220 21

42 26

88 15

10 16

142 77

63 64

32 32

29 37

120 35

43 40

QS SA

REM

QS GR

REM

QS vo

REM

QS TO

REM

QS Mean

REM

&s

CV, coeff of variation (SD as percentage of mean); MVR, minute ventilatory response (dvI/dPETco,); MIF, mean inspiratory flow [(d(VT/TI)/dP ETCH,)]. See Table 2 for further details.

172

RESPONSE

OF THE

NEWBORN

TO CO, DURING

SLEEP

OS QS REM &$t) and mean inspiratory flow responses (rig&) for REM sleep and QS for each of 5 subjects. Each point is mean of 4-5 tests (for individual tests see Fig. 1). REM sleep and QS responses for each subject are joined. Note consistently reduced response during REM sleep.

REM

FIG. 3. Ventilatory

VWTI during rebreathing and the lower correlation coefficients for regression analyses. These findings are consistent with the increased

irregularity

and variability

of

ventilation during REM sleep in the newborn (3) and with the reportedresponse of infants (16), adults (2), and dogs (25) during CO, rebreathing. The behavioral features of REM sleep (e.g., twitching and eye movements), which are manifestations of cortical and reticular activation, are thought to underlie the irregularity of CO, responses in REM sleep. The response during QS is probably more regular because of the dominant role of chemoreceptor and vagal afferents and the relative absence of a behavioral component (24, 31). Short-term reproducibility of the ventilatory response

to CO, by use of the rebreathing method has been reported for awake adults (1, 30). Adults continue to exhibit considerable variability during sleep, although coefficients of variation have not been specifically reported (2,lO). Values for the newborn have not been reported. We found mean coefficients of variation during QS to be comparable to the maximum values reported for awake adults studied with the rebreathing technique (34%) (l), with considerably greater variation during REM sleep. The general trend in the variability of the ventilatory response between REM sleep and QS is therefore generally consistent with what might be expected on the basis of examination of breath-to-breath variability of individual responses. Interestingly, ventilatory responses dur-

l

120

l

3

CT n

90

-

3%

l

&z c,

50

55 End-Tidei CO2 ImmHgl

60

65

] s;,*

60

O45

*

l

:

50

55 60 End-Tidal CO2 tmmHg1

FIG. 4. Breath-by-breath changes in 01 (left) atid VT/TI (right) during REM sleep test (subject SA). Note absence of correlation between PETE*, and VI (r = 0.00) but significant correlation betweenP~~~~, and VT/TI (r = 0.41, P < 0.002). (Lack of ventiiatory response was due to significant negative slope of f response.)

65

RESPONSE

OF THE

NEWBORN

ing REM sleep and QS were sometimes similar, although they diverged widely at other times (Fig. 2). In the dog, ventilatory response to inhaled CO, has been reported to be brisk during the “tonic” phase of REM sleep but is seemingly dissociated from the CO, stimulus during bursts of phasic activity (32). We found it difficult (and therefore did not attempt) to separate the tonic and phasic components of REM sleep during our studies; the variable REM sleep response may well reflect the influence of a variable phasic component. Occasional deviations from the general trend of increased variability in REM sleep vs. QS were shown, by one subject (TO) with low REM sleep variability and another (VO) with comparatively high QS variability (Table 3). This suggests that, regardless of apparent outward manifestations of regularity and stability, there should not be unrealistic assumptions made about the reproducibility of stimuliresponse relationships during QS. In fact, this should be anticipated because it is unlikely that there is a stable physiological steady state during either QS or REM sleep in the newborn (29). We found no evidence indicating that the relatively modest changes in Ptco, that occurred during rebreathing were correlated with test-to-test differences in the ventilatory response. However, fluctuations in arterial PO, may elicit chemoreflex-mediated changes in ventilation (18). It would be of interest to compare the variability we report with that measured during hyperoxic rebreathing, when presumably any peripheral chemoreceptor activity is minimized; such data have not been published. Honma et al. (16) concluded that interindividual variability in the slope of the CO, response was due to differences in the VT and f contributions to increased ventilation during QS, whereas during REM sleep it was due principally to the marked fall in f. Our findings extend these observations to indicate that the same factors are responsible for short-term within-subject variability. The prominent slowing of f during CO, inhalation has previously been reported during CO, rebreathing in the dog (25), adults (9), and the newborn (16). It may be a reflex response to state-related changes in the resistance or compliance of the respiratory system (11, 17, 26), vaQS

REM

TO CO2 DURING

173

SLEEP

TABLE 4. Summary of uentilatory

changes State

REM Ptcoz, Torr diTI/dPET,,,

76.4~~0.7

75.7t0.6