OBJECTIVE:To determine whether inhalation of 0.8% CO(2) in preterm infants decreases the duration and rate of apnea as effectively as or better than theophylline with fewer adverse side effects. STUDY DESIGN:A prospective, randomized, control study of 42 preterm infants of gestational age 27 to 32 weeks assigned to receive inhaled CO(2) (n = 21) or theophylline (n = 21). The study group had a mean (+/- standard error of the mean) birth weight of 1437 +/- 57 g, gestational age of 29.4 +/- 0.3 weeks, and postnatal age of 43 +/- 4 days. After a control period, 0.8% CO(2) or theophylline was given for 2 hours, followed by a recovery period. RESULTS:In the CO(2) group, apneic time and rate decreased significantly, from 9.4 +/- 1.6 seconds/minute and 94 +/- 15 apneic episodes/hour to 3.0 +/- 0.5 seconds/minute and 34 +/- 5 apneic episodes/hour. In the theophylline group, apneic time and rate decreased significantly, from 8 +/- 1 seconds/minute and 80 +/- 8 apneic episodes/hour to 2.5 +/- 0.4 seconds/minute and 28 +/- 3 apneic episodes/hour. Cerebral blood flow velocity (CBFV) decreased only during theophylline administration. CONCLUSIONS:Our findings suggest that inhaled low (0.8%) CO(2) concentrations in preterm infants is at least as effective as theophylline in decreasing the duration and number of apneic episodes, has fewer side effects, and causes no changes in CBFV. We speculate that CO(2) may be a better treatment for apnea of prematurity than methylxanthines.

We measured the PCO2 apneic threshold in preterm and term infants. We hypothesized that, compared with adult subjects, the PCO2 apneic threshold in neonates is very close to the eupneic PCO2, likely facilitating the appearance of periodic breathing and apnea. In contrast with adults, who need to be artificially hyperventilated to switch from regular to periodic breathing, neonates do this spontaneously. We therefore measured the apneic threshold as the average alveolar PCO2 (PaCO2) of the last three breaths of regular breathing preceding the first apnea of an epoch of periodic breathing. We also measured the PaCO2 of the first three breaths of regular breathing after the last apnea of the same periodic breathing epoch. In preterm infants, eupneic PaCO2 was 38.6 +/- 1.4 Torr, the preperiodic PaCO2 apneic threshold was 37.3 +/- 1.4 Torr, and the postperiodic PaCO2 was 37.2 +/- 1.4 Torr. In term infants, the eupneic PaCO2 was 39.7 +/- 1.1 Torr, the preperiodic PaCO2 apneic threshold was 38.7 +/- 1.0 Torr, and the postperiodic value was 37.9 +/- 1.2 Torr. This means that the PaCO2 apneic thresholds were 1.3 +/- 0.1 and 1.0 +/- 0.2 Torr below eupneic PaCO2 in preterm and term infants, respectively. The transition from eupneic PaCO2 to PaCO2 apneic threshold preceding periodic breathing was accompanied by a minor and nonsignificant increase in ventilation, primarily related to a slight increase in frequency. The findings suggest that neonates breathe very close to their PCO2 apneic threshold, the overall average eupneic PCO2 being only 1.15 +/- 0.2 Torr (0.95-1.79, 95% confidence interval) above the apneic threshold. This value is much lower than that reported for adult subjects (3.5 +/- 0.4 Torr). We speculate that this closeness of eupneic and apneic PCO2 thresholds confers great vulnerability to the respiratory control system in neonates, because minor oscillations in breathing may bring eupneic PCO2 below threshold, causing apnea.

Periodic breathing and apnea are common in neonates, yet the physiological mechanisms involved are not clear. A low arterial PO2 might magnify peripheral chemoreceptor contribution to breathing, with its baseline variability inducing major changes in ventilation, leading to instability of the respiratory control system. We hypothesized that neonates: (1) would depend much more on the peripheral chemoreceptor contribution to breathing than adult subjects and (2) their baseline arterial PO2 would sit on the steep portion of the ventilation/arterial PO2 relationship on the adult nomogram, making breathing prone to oscillate. We analyzed data from previous polygraphic recordings in four groups of subjects: small preterm infants [SPI; postconceptional age (PCA) 33 ± 2 weeks; n = 40], large preterm infants (LPI; PCA 36 ± 2 weeks; n = 34), term infants (TI; PCA 42 ± 1 week; n = 24), and adult subjects (AS; weight 63 ± 2 kg; age 29 ± 3 years, n = 16). Peripheral chemoreceptor activity was measured by: (1) the immediate decrease in ventilation and (2) apnea time during brief inhalation of 100% O2 (about 1 minute). We found that: (1) the immediate decrease in ventilation with 100% O2 was more pronounced in infants than in adult subjects (38 ± 2 versus 6 ± 5%), and in infants breathing periodically versus those breathing continuously; (2) the apnea time during 100% O2 was also significantly longer in periodic breathing infants; and (3) the TcPO2 was much lower in infants than in adult subjects (65 ± 1 versus 93 ± 1 Torr), and also lower in periodic versus continuously breathing infants. It was located significantly to the left of values for the adult subject, on the ventilation/arterial PO2 diagram. The data suggest that: (1) a substantial portion of baseline breathing activity early in life is maintained by increased peripheral chemoreceptor activity; and (2) neonates breathe irregularly with apneas due to the position of their arterial PO2 values on the ventilation/arterial PO2 diagram, in which a change in PO2 produces a more significant change in ventilation than that observed later in life.

To test the hypothesis that the crescendo–decrescendo type of pattern of periodic breathing is more common in infants than in adulthood, we examined the morphologies of periodic breathing in four groups of subjects: group 1 (n=10, gestational age 30±1 week), group 2 (n=10, GA 31±1 week), group 3 (n=10, GA 38±1 week), and group 4 (n=10, age 50±4 years). Respiratory pattern and ventilation were measured using a flow-through system. The breathing morphologies were defined according to the respiratory flow. We found (1) a predominant crescendo–decrescendo pattern in preterm infants (groups 1 and 2, >50%) and this changed to a predominant decrescendo breathing in adults (group 4, 50%); (2) total breathing cycle and its phases did not change significantly among the neonatal groups, but they almost doubled in adult subjects; however, the number of breaths per breathing interval remained the same (crescendo–decrescendo) or less (flat and decrescendo) in adults as compared to preterm infants; (3) the duty cycle (breathing interval/cycle duration) remained consistent with age; and (4) at the beginning of each breathing interval, alveolar Pco2 was highest and alveolar Po2 and O2 saturation lowest. The findings suggest a change in the strategy of the respiratory control system during periodic breathing between the infant and the adult, perhaps dictated by mechanical and chemoreceptor limitations early in age, with a switch from a crescendo–decrescendo to a predominantly decrescendo pattern.

We previously reported on the presence of respiratory pacemaker cells that are highly sensitive to CO(2), in a region of the medulla oblongata in the fetal rat, 2 mm rostral to the obex. We now report on the CO(2) dose responses of these cells, as well as their responsiveness to certain chemical agents known to affect breathing in the fetus. Twenty-day-old fetal Sprague Dawley rats were block-dissected, and the cells of target areas were dissociated as previously described. Neuronal cells were plated on a medullary background and placed in the incubator with 10% CO(2) for 2-3 weeks. Cells were then studied using patch-clamp techniques. Pacemaker cells with single or bursting potentials showed responsiveness to CO(2) starting with pulses of 10 msec. Irregular beating or silent cells had poor or absent responsiveness to CO(2). Pacemaker cells responded to norepinephrine with increased firing potential; this action was blocked by metropolol. PGE(2) had no effect on pacemaker-cell activity, but indomethacin increased the spike frequency from 336+/-41 to 384+/- 65 spikes/min. Morphine stimulated the pacemaker cells from 205+/-25 to 272+/-29 spikes/min; this was blocked by naloxone. Finally, a placental extract, which inhibited breathing in the unanesthetized fetal sheep preparation, increased the activity of pacemaker cells from 301+/-35 to 452+/-52 spikes/min. In all of the above, irregular beating cells responded poorly and silent cells did not respond. The findings indicate that these pacemaker cells are uniquely designed to respond to CO(2) and have some properties which allow them to respond to certain chemical mediators in a manner similar to that of the whole respiratory system in vivo.

To examine the influence of sleep state, respiratory pattern, and ventilation on cyclical fluctuations (CF) in cerebral blood flow (CBF) velocity (CBFV), we studied 21 'healthy' preterm infants: birth weight 1,790 +/- 162 g (SEM), study weight 1,960 +/- 165 g, gestational age 32 +/- 1 weeks, postnatal age 20 +/- 4 (range 8-57) days. The CBFV was measured using on-line pulsed Doppler ultrasound by insonating the middle cerebral artery. Breathing was measured using a flow through system. The sleep state was monitored according to conventional criteria. Three hundred and seventy-five epochs of 1 min each were analyzed; 207 during quiet sleep (QS) and 168 during rapid eye movement (REM) sleep. CFs in CBFV were detected in all babies. The frequency of CF ranged from 0.5 to 6 cycles/min. The proportion of epochs showing CF was similar during both sleep states (56% QS vs. 59% REM; p = NS). Although the mean CBFV (cm/s) was similar in these two sleep states, the mean coefficient of variation, a measure of CF amplitude, was significantly higher during REM as compared with QS (6 +/- 0.5 vs. 4.3 +/- 0.2%; p < 0.05). Similarly, the mean CBFVs were similar with various respiratory patterns, but the coefficient of variation was significantly higher in periodic and apneic patterns as compared with regular and irregular respiratory patterns (5.6 +/- 0.6% periodic, 5.6 +/- 0.3% apneic, 3.6 +/- 0.3% regular, and 4.1 +/- 0.5% irregular, p < 0.05). The amplitude of CF was associated with the variability of the heart rate (p < 0.05), but not with the variability of the respiratory measurements. These findings suggest: (1) REM sleep is associated with a greater CBF variability than QS, and (2) periodic and apneic breathing are associated with a greater CBF variability than regular or irregular breathing. We speculate that sleep state and respiratory pattern do not determine but modulate the CBF. Our data suggest that in studies involving interpretation of CBFV data using the Doppler technique, breathing patterns should be taken into account in addition to sleep state.

We have shown previously that administration of 100% O2 to preterm infants induces an apnea which in about 20% of cases has an obstructive component. The obstruction occurred during the longer apneas. In the present study, we tested the hypothesis that the appearance of obstruction in longer apneas depends on the baseline oxygenation. Sixteen preterm infants were studied in quiet sleep (birthweight 1435 +/- 93 g [mean +/- SEM], study weight 1711 +/- 90 g, gestational age 30 +/- 1 weeks, and postnatal age 26 +/- 5 days) at various baseline oxygenations. A flow-through system was used to measure ventilation and alveolar gases. Respiratory efforts in the absence of flow were detected using chest and abdominal displacements or diaphragmatic electromyography. Each infant inhaled 15%, 21%, or 25% O2 for 5 minutes (control period) followed by 100% O, for 2 minutes. Baseline alveolar PO2, O2 saturation and transcutaneous PO2 increased during inhalation of 15%, 21%, and 25% O2, respectively. The immediate decrease in ventilation with 100% O2 was 52% on 15% O2, 20% on 21% O2, and 16% on 25% O2 (p < 0.001); this was associated with an apnea in all cases. The mean length of apneas during 100% O2 was 37 seconds on 15% O2, 19 seconds on 21%, and 11 seconds on 25% (p < 0.01). Twelve infants (75%) developed mixed obstructive apneas in response to 100% O2 when breathing 15% O2 during control period, three (19%) when breathing 21% O2, and none had mixed apnea when breathing 25% O2 during control period (p < 0.05). These findings suggest that lower baseline oxygenation predisposes to long mixed apneas. We speculate that the peripheral chemoreceptors, uniquely active in the small preterm infant with relatively low O2 tension, when suppressed by an increase in oxygen tension, trigger a central inhibition with loss of upper airway muscle tone. This is more pronounced when the baseline oxygen tension is lower, leading to more prolonged apnea and increased probability of obstruction.

Oral breathing is an important defense mechanism, yet its prevalence and relationship to behavioral activities have not been studied in preterm infants. We tested the hypothesis that oral breathing is rare in these infants and likely to be restricted to periods of body movements. Ten healthy preterm infants (birthweight 1300 +/- 100 g [SE]; gestational age 29 +/- 1 weeks; postnatal age 36 +/- 7 days) were studied. Ventilation was measured with a nose piece and screen flowmeter. Oral breathing was detected with a carbon dioxide sampler at the mouth. Movements were classified according to intensity into type I (localized, minor signal distortion) and type II (generalized, moderate signal distortion). Oral breathing was present 10% of the time, with a mean duration of 27 +/- 3 seconds. Of 104 episodes of oral breathing, 13 (13%) occurred during type I movement, 89 (86%; p < 0.01) during type II, and 2 (2%) in the absence of movement. The delay from beginning of movements to the beginning of oral breathing was 20 +/- 3 seconds. Nasal minute ventilation decreased from 0.203 +/- 0.013 L.min-1.kg-1 during movements in the absence of oral breathing to 0.167 +/- 0.013 L.min-1.kg-1 during movements plus oral breathing (p = 0.017). In 496 type I and II movements, the prevalence of oral breathing was 21 of 165 (13%) in quiet sleep, 37 of 194 (19%) in rapid eye movement sleep, 6 of 12 (50%) in transitional sleep, and 44 of 125 (35%) in indeterminate sleep (p < 0.01).(ABSTRACT TRUNCATED AT 250 WORDS)