Pulmonary interstitial emphysema

SUMMARY Forty one of 210 preterm infants ventilated for respiratory distress syndrome in a three year period had radiological evidence of pulmonary interstitial emphysema. The development of this condition was significantly associated with malpositioning of the endo-tracheal tube in a main bronchus and the use of high peak pressure ventilation. Pulmonary interstitial emphysema was associated with a significant increase in the number of pneumothor-aces, intraventricular haemorrhages, and the need for prolonged respiratory support, but did not increase mortality. Although in 12 infants in whom fast rate ventilation was used there was a significant reduction in the number of pneumothoraces, outcome was not altered in any other way. Fast rate ventilation may be of greater benefit if initiated before the development of pulmonary interstitial emphysema.


ventilation.6
We have studied the incidence, aetiological as- sociations, and morbidity of pulmonary interstitial emphysema in preterm babies in the Neonatal Intensive Care Unit in Cambridge ventilated for respiratory distress syndrome during the years 1980-83, and have assessed the effect of fast rate ventilation on the morbidity and mortality of this condition.

Patients and methods
We reviewed retrospectively the notes and chest radiograph reports of 254 babies ventilated on the unit during the three year period August 1980 to July 1983.Eight infants with fatal congenital abnormalities (Potter's syndrome and chromosomal abnormalities) were immediately excluded from the study.A total of 210 babies of gestational age 24 to 35 weeks were ventilated because of increasing respiratory failure due to the respiratory distress syndrome.Respiratory distress syndrome was di- agnosed on the basis of classic clinical features: cyanosis in air, retractions and an expiratory grunt developing in the first hours of life and persisting for at least 48 hours, and a chest radiograph showing an air bronchogram and diffuse atelectasis.Intraventricular haemorrhage was diagnosed by real time ultrasound scan; only haemorrhages greater in size than a subependymal bleed are reported in the present study.
The original chest radiograph reports had been issued by a radiologist who had only limited knowledge of the infant's clinical condition (essentially 'blind').Those radiographs showing pulmonary interstitial emphysema, as described by Campbell, were subsequently reviewed.The timing of the development of the condition was noted and a careful assessment of endotracheal tube siting was made.The first chest radiograph to show evidence and that showing the worst pulmonary interstitial emphysema were further analysed.A scoring system was devised whereby each of the three zones (upper, middle, and lower) of each lung was subjectively assessed and given a score 0 to 9, taking into account both the severity (graded 1 to 3) and the distribution of the pulmonary interstitial emphysema (graded 1 1046 This was the first radiograph to show pulmonary interstitial emphysema.Severe (grade 3) centrally placed (grade 1) interstitial air scored 3 points in each midzone.making a total of 6 points for both lungs.There is also a left basal pneumothorax.The most noticeable changes were seen on day 4.This radiograph scored 24. with pulmonary interstitial emphysema only really present in the right lung (9, 9, 9).On no radiograph was the endotracheal tube seen to lie in the right main bronchus.The highest scoring radiograph (45) was seen on day 3.At this stage the right lung scored 18 (6.9, 3) while the left scored 27 (9. 9. 9).
to 3).These latter factors were multiplied together so that mild disease, confined to just the medial portion of the zone (grade 1), would only score one point, whereas severe (grade 3) disease extending right out to the periphery of the lung (grade 3) would score 9 points.The totals for each lung were recorded (maximum 27 points) and any asymmetry between the two lungs was noted.The total for each chest radiograph was calculated (maximum 54 points).Representative examples are shown in Figs.
1 to 4. Using this scoring system we determined if the mortality and morbidity were related to the severity of the interstitial emphysema.
Statistical tests.x2 test and the Student's t test were used for statistical analysis.

Results
Aetiological associations.Forty one of the 210 infants ventilated for respiratory distress syndrome developed radiological evidence of pulmonary in- terstitial emphysema.The incidence was not related to gestational age (Table 1) or birthweight, and although it was more common in babies resuscitated at birth by positive pressure ventilation through an endotracheal tube, this did not reach statistical significance.Nor was there a significant increase among babies ventilated from birth.
Fourteen infants developed asymmetrical pul- monary interstitial emphysema, one lung being at least twice as severely affected as the other.Such asymmetry occurred in seven of the only 10 infants with malpositioned tubes (P<0.05).Among the six infants with true unilateral pulmonary interstitial emphysema, however, only one had had a malpositioned endotracheal tube.
The most significant aetiological association was the maximum peak pressure that the infant had been exposed to before the condition developed (Fig. 5).
This was significantly higher in those infants de- veloping pulmonary interstitial emphysema com- pared with those who did not (P<0.001).There was no significant difference, however, in the maximum peak pressure in those infants who  Morbidity and mortality.There was a significant increase in the incidence of pneumothoraces among babies with pulmonary interstitial emphysema com- pared to those without it-31 of 41 v 41 of 169 (P<0.001).The development of a pneumothorax, however, was not correlated with the severity of the pre-existing condition, as assessed by the scoring system.The maximum peak pressure that the infant was exposed to before the development of pulmonary interstitial emphysema was the same in those infants who did not develop further air leaks as in those who subsequently developed a pneumothorax.Twenty one of the 41 babies with pulmonary interstitial emphysema developed an intraventricu- lar haemorrhage compared with an incidence of 39 of 169 among the remainder (P<0.001).Intraventricular haemorrhages were also more com- mon in infants with pulmonary interstitial emphysema who did not subsequently develop pneumothoraces (4 of 10) compared with infants with respiratory distress syndrome who had no form of air leak (18 of 128).Infants with intraventricular haemorrhages tended to have more severe disease, as judged by the scoring system, but this did not reach statistical significance.
The development of pulmonary interstitial emphysema increased the number of babies requir- ing ventilation for longer than 168 hours or sup- plementary oxygen for more than 240 hours com- pared with the remainder (P<0.05), (  pulmonary interstitial emphysema were compared with babies without any form of air leak, respiratory distress syndrome alone, (P<0001), (Table 2).Pulmonary interstitial emphysema without a subse- quent pneumothorax also significantly increased the number of infants requiring prolonged oxygen ther- apy (P<0001) and artificial ventilation (P<0-01) (Table 2) compared with the respiratory distress syndrome alone group.Among the infants with pulmonary interstitial emphysema, however, the addition of a further air leak did not alter the need for prolonged respiratory support (Table 3).There was, however, no significant difference in the number of infants with and without pulmonary interstitial emphysema developing chronic lung dis- ease (defined as the need for more than three weeks respiratory support).
The mortality of babies with pulmonary intersti- tial emphysema (10 of 41) was not significantly higher than that of infants without the condition (31 of 169) and there was no statistical difference in the mortality rates between infants with pulmonary interstitial emphysema who subsequently developed pneumothoraces and those who did not (Table 4).Infants with this condition who died tended to have more severe disease, as assessed from their chest radiographs but this did not reach significance.
Treatment.Twelve infants were treated with fast rate ventilation (rates greater than 110/minute) as soon as pulmonary interstitial emphysema was diagnosed.The infants in both treatment groups, that is receiving slow or fast rate ventilation, were of similar gestational ages and postnatal ages when the condition was first diagnosed (Table 5).Although the infants started on fast rate ventilation tended to have more severe disease this did not reach statistic- al significance.Nine infants were paralysed as soon as they developed pulmonary interstitial emphysema (five from the fast rate group) and two subsequently developed a pneumothorax.All four babies paralysed on slow frequency ventilation developed further air leaks.
In eight infants the change to fast rate ventilation enabled peak pressures to be lowered within the subsequent two hours (mean 4-9 cm H20, range 2 to 8 cm H20), however, in two others peak pressures had to be increased (mean 2-5 cm H20, range 2 to 3 cm H20).
Only two infants treated with fast rate ventilation subsequently developed a pneumothorax, one of whom had required increased pressures on changing to the faster rates.Twenty nine babies remained on conventional ventilator rates (rate 30 to 40) and 25 developed pneumothoraces (P<0001).
Although fast rate ventilation reduced the inci- dence of penumothoraces, the pulmonary interstitial emphysema in those infants deteriorated significantly (P<005) (Table 5).There was less progression of the disease in babies remaining on slow rate ventilation, however most had developed pneumothoraces.Despite a reduction in the number of pneumothoraces in the fast rate group, there was no other improvement in outcome; in particular no significant alteration in the mortality, incidence of in- traventricular haemorrhage, or requirement for prolonged respiratory support.

Discussion
Pulmonary interstitial emphysema may occasionally occur spontaneously in preterm babies with respiratory distress syndrome, but it is more usually associated with the use of intermittent positive pressure ventilation8 and may be another consequ- ence of iatrogenic barotrauma.In this study the most significant association with the development of this condition was the use of high peak pressures, as has been found with pneumothoraces.9The in- creased severity of disease found after incorrect positioning of the endotracheal tube down one main bronchus, seen in this study and others,8 is further evidence for an aetiological role of barotrauma in this condition.Very preterm infants exposed to such high inflating pressures are more likely to develop pulmonary interstitial emphysema, as the increased amount of pulmonary connective tissue in the immature lung") is more likely to trap air which has ruptured an alveolus.
The mortality in this study (24%) compares favourably with the 67% 11 and 39% 12 reported previously.We were unable to confirm that the presence of pulmonary interstitial emphysema significantly increased the mortality nor, as reported by Gregoire,12 that where it developed on the first postnatal day it was always bilateral and usually fatal.
We were able to show that fast rate ventilation significantly reduced the number of penumothoraces among these infants.The use of lower peak pressures used in most infants at fast rate ventilation may in part explain this reduced incidence.In other respects the final outcome of these infants was not altered, there being no difference in the mortality, incidence of intraventricular haemorrhage, or chro- nic lung disease between the fast or slow rate group; however, the mortality in our series was already low." 12The failure to reduce the incidence of intraventricular haemorrhage and chronic lung dis- ease in the group undergoing fast rate ventilation suggests that pulmonary interstitial emphysema alone is an important cause of morbidity and its prevention is therefore as important as preventing further air leaks.
It is interesting to speculate why the pulmonary interstitial emphysema actually worsened during fast ventilation.In the absence of a pneumothorax, and particularly a pneumomediastinum, the pulmonary interstitial emphysema would be unable to decompress.8At the faster ventilator rates insuffi- cient expiratory time could have caused further accumulation of trapped air.Inadvertent positive end expiratory pressure can be a consequence of fast rates and could have compounded this problem, but it was not encountered in this study.
We have been able to show that the use of high peak pressure ventilation is important in the causa- tion of pulmonary interstitial emphysema and that fast rate ventilation may prevent pneumothoraces from occurring subsequently.In this study, howev- er, fast rate ventilation was only used after the infants had already been exposed to high pressure ventilation, which may explain the relative lack of success of this form of treatment in reducing serious morbidity and mortality.We would therefore sug- gest from these preliminary findings that a control- led study of the use of fast rate ventilation from birth before the pulmonary interstitial emphysema has developed is necessary.

Fig. 2
Fig. 2 Chest radiograph showing pulmonary interstitial emphysema in an infant of25 weeks' gestation.This infant developed minimal signs on the first postnatal day: the most noticeable changes were present on day 3 and are shown in the figure.This radiograph scored a total of 13 points-the right lung 10 (4. 4, 2) and the left 3(1, 1, 1).

Fig. 5
Fig. 5 Relation between peak ventilation pressures and the development ofan air leak in infants ventilated for respiratory distress syndrome.PIE=pulmonary interstitial emphysema.PTX= pneumothorax

Table 1
Gestational age distribution ofthe 210 preterm infants with respiratory distress sydrome with or without air

Table 2 )
, and was especially significant if the 41 infants with Arch Dis Child: first published as 10.1136/adc.59.11.1046 on 1 November 1984.Downloaded from

Table 2
Incidence ofprolonged respiratory support in preterm infants with respiratory distress syndrome with or without

Table 3
Incidence ofneedfor prolonged respiratory support in preterm infants with pulmonary interstitial emphysema

Table 4
Mortality among 210 preterm infants with respiratory distress syndrome with or without air leaks

Table 5
Clinical characteristics of two groups ofpreterm infants with respiratory distress syndrome and pulmonary interstitial emphysema treated with fast rate or slow rate ventilation. (Values mean (SD))