Oxygenation assessed noninvasively during sleep, feeding and wakefulness was normal in all infants at 6 months of age. Although 67 percent of ECMO survivors experienced short periods of apnea of 4- 10-s duration, none of these episodes was clinically significant. No episode of severe desaturation less than 90 percent (equivalent to Pa02 of 60 mm Hg) was observed, as have been previously described in infants with chronic lung disease. These findings may suggest less severe lung disease in ECMO survivors than in infants with BPD following prematurity. This is further supported by the fact that the Cdyn measurements in ECMO survivors were significantly better than those of preterm BPD infants. Swaminathan et al have reported that long-term pulmonary function abnormalities following meconium aspiration syndrome in term infants are less severe than those following BPD in preterm infants. Since ECMO patients are, by definition, near-term infants, one would expect residual chronic lung disease to be milder than that seen in preterm BPD survivors.
It is commonly believed that the avoidance of further injury from treatment, using lung rest on ECMO, will allow lung healing and recovery. However, the results from this study suggest that these infants have such severe initial lung injury that chronic lung disease cannot be completely prevented by this mode of therapy.
However, only one of the infants required rehospitalization during the first six months for respiratory deterioration. Additionally, chronic lung disease following ECMO was less severe in this group when compared with that in preterm BPD infants at a similar age. The most common abnormalities were increased Raw and decreased SGaw. These findings were further confirmed by the fact that the most commonly required medication was a bronchodilator (52 percent). Therefore, ECMO and lung rest may modify the severity of chronic lung disease, but do not appear to prevent it altogether, as previously suggested.
Due to lack of a control population, the pulmonary mechanics data from ECMO survivors were compared with those of 13 preterm BPD infants, survivors of severe RDS, studied at 6 months of corrected age. These data have been reported previously. The mean Cdyn in the BPD group was 0.52 + 0.06 ml/cm HaO which is significantly lower than that of the ECMO survivors (p<0.001 [Fig 2]). The TGV, Raw and P&02 were comparable in the two groups at 6 months corrected age (Fig 2).
Dynamic pulmonary compliance at 6 months of age did not correlate with the duration of ECMO (r = 0.23; NS) or the duration of mechanical assisted ventilation required following ECMO (r = 0.17; NS). There was no significant correlation between the pulmonary function abnormalities at 6 months of age and the severity of respiratory failure pre-EC MO, such as pre ECMO Pa02, OI, MAP and the age at starting ECMO.
Feeding was recorded in 15 infants (11.3 ± 1.8 min) and awake time recording was obtained in 16 infants (16.5 ±2.4 min). The mean observed Sa02 during wakefulness (98.7 ±0.3) and feeding (99 ±0.2) was comparable to the same value during sleep (98.7 ±0.3 percent) (NS). No significant desaturation (Sa02 <90 percent) was observed during sleep. Only two infants (11 percent) experienced brief episodes of desaturation for 5 s during feeding and wakefulness. Twelve infants (67 percent) experienced short periods of apnea of 4- 10s duration during sleep. These apneic episodes were not associated with bradycardia or oxygen desaturation. There was no significant difference in the average number of apneic episodes observed during active sleep or quiet sleep.
Nine infants (47 percent) required supplemental oxygen for part or all of the first six months of life because their Pa02 values were less than 70 mm Hg on FIo2 0.21 percent. At 6 months of age only one patient still required supplemental oxygen and this was the only infant that required rehospitalization during the first six months for respiratory failure. Three other infants had upper respiratory tract infections and otitis media during these six months but they were managed on an outpatient basis. Arterial blood gas levels by radial artery puncture were obtained in 12 infants at 6 months of age. Transcutaneous oxygen tension was monitored continuously during arterial puncture for correlation. The mean Pa02 was 80 ±5 mm Hg. Five of 12 infants (42 percent) still had Pa02 <80 mm Hg while they were breathing room air (Fig 1). The mean pH was 7.44 ±0.01 and PaC02 was 31 ±2 mm Hg, representing acute hyperventilation secondary to crying during arterial puncture.
Prior to ECMO all infants demonstrated severe gas exchange abnormalities, with the mean lowest Pa02 being 32 ±3 mm Hg while on maximal mechanical ventilator settings such as FIo2 of 1.0 and the MAP being 18.7 +1.2 cm H20 (Table 1). One infant had cardiac arrest requiring resuscitation prior to ECMO. The mean highest OI was 79 ±20, and therefore all study subjects met the institutional criteria of a greater than 80 percent mortality on conventional therapy (OI >40) prior to ECMO.
An esophageal balloon (1.0 cm in diameter and 4.0 cm long) containing 0.2 ml of air was passed until it was positioned in the lower third of the esophagus, according to the method of Beardsmore et al. The Cdyn was then measured. The maximal intrasubject coefficient of variation for Cdyn in our laboratory is 8.7 ±4.3 percent.
Unless otherwise specified, all data are expressed as group means and standard errors of the mean. The relationship between variables was expressed as the correlation coefficient using linear regression analysis by the least squares method.
Minute ventilation, Vt and respiratory frequency were measured during an average of at least 20 breaths during quiet breathing using an oronasal mask. The door of the body plethysmograph was closed and the infant breathed air from the box. When thermal equilibrium had been achieved, the pneumatic sleeves were sealed. An infant oronasal mask connected to a rebreathing bag (inspired air at body temperature and pressure saturated) and a heated pneumotachograph with a shutter was sealed over the infants nose and mouth. The mask was checked for leakage both visually and by appearance of the oscilloscope tracings.
The heart rates determined from the ECG and the pulse oximeter were compared and Sa02 signals were rejected as artifacts if the difference was greater than 5 beats per minute or damping of the pulse tracing from the oximeter was seen. Wakefulness, sleep state and feeding were determined by direct observation and behavorial criteria. An arterial blood gas level was obtained at the end of the sleep study by radial artery puncture and compared with the Sa02 value and tc02.