Published ahead of print on December 23, 2003, doi:10.1164/rccm.200310-1425OC
© 2004 American Thoracic Society
Pulmonary Function at Follow-up of Very Preterm Infants from the United Kingdom Oscillation StudyDepartment of Child Health, Guy's King's & St. Thomas' Medical School, King's College Hospital; Department of Community Health Sciences; Department of Child Health, St. George's Hospital Medical School, London; and Department of Child Health, University Hospital, Nottingham, United Kingdom Correspondence and requests for reprints should be addressed to Anne Greenough, M.D., M.B.B.S., D.C.H., F.R.C.P., F.R.C.P.C.H., Department of Child Health, 4th floor Golden Jubilee Wing, King's College Hospital, London SE5 9RS, UK. E-mail: anne.greenough{at}kcl.ac.uk
Prematurely born infants supported by conventional ventilation (CV) frequently have abnormal pulmonary function when assessed in childhood. The aim of this study was to test the hypothesis that infants who were randomly assigned to high-frequency oscillatory ventilation would have superior pulmonary function at follow-up compared with those who received CV (UK Oscillation Study). Infants from 12 trial centers were recruited for pulmonary function testing at a single center. Seventy-six infants, of a mean gestational age 26.4 weeks, were studied after sedation with chloral hydrate at between 11 and 14 months of age, corrected for prematurity. Infants assigned to CV had similar pulmonary function compared with those assigned to high-frequency oscillatory ventilation, with mean (SD) results as follows: functional residual capacity measured by whole-body plethysmography, 26.9 (6.3) versus 26.5 (6.4) ml/kg; functional residual capacity measured by helium dilution, 24.1 (5.4) versus 23.5 (5.7) ml/kg; inspiratory airway resistance, 3.3 (1.3) versus 3.4 (1.6) kPa · second · L; expiratory airway resistance, 4.4 (2.8) versus 4.1 (2.5) kPa · second · L; respiratory rate, 31.2 (6.0) versus 33.9 (8.0) breaths/minute. We conclude that early use of high-frequency oscillatory ventilation in very preterm infants appears to offer no advantage over CV in terms of pulmonary function at follow-up.
Key Words: high-frequency ventilation • neonatal chronic lung disease • lung volume measurements • airway resistance Respiratory morbidity remains a major outcome after very preterm birth. This is shown at follow-up by a high incidence of respiratory symptoms, frequent hospitalization for respiratory illness, and abnormal pulmonary function (1–4). A number of strategies have been employed in the neonatal period with the aim of reducing chronic lung disease and hence such sequelae, including ventilation techniques designed to avoid pulmonary volutrauma. High-frequency oscillatory ventilation (HFOV) showed initial promise as an effective strategy for reducing pulmonary complications (5–8), but recent randomized controlled trials studying its use as an elective mode of ventilation early in the postnatal course of very preterm infants have shown little or no benefit in short-term respiratory outcome (9–11). The UK Oscillation Study (UKOS) has been the largest randomized trial to date and compared HFOV with conventional ventilation (CV) as the initial ventilatory modality in preterm infants on admission for neonatal intensive care (12). Although there were similar rates of chronic lung disease, defined as oxygen dependency at 36 weeks postmenstrual age, in each ventilator group, we remained concerned that this relatively unsophisticated marker of respiratory outcome might not detect more subtle differences between the groups. These potential differences could become apparent as the infants grew older, as recent evidence demonstrates deterioration in airway function during the 1st year of life in very preterm infants, regardless of lung disease severity (13, 14). The only previous randomized study to include more detailed respiratory follow-up and measurement of pulmonary function in infancy was the High Frequency Ventilation in Premature Infants Study (HIFI) (15, 16). Infants in this trial, however, were relatively mature and did not receive antenatal steroids or exogenous surfactant. More importantly, no strategies to optimize lung volume on HFOV were employed (17). No differences in lung function in infancy were found, but these results cannot be applied to the current population of very preterm infants, who are treated with early HFOV using a lung volume optimization strategy after antenatal steroids and exogenous surfactant therapy. Any benefits or adverse effects conferred by HFOV beyond the neonatal period, therefore, remained unclear. The aim of this study was to test the hypothesis that infants who had been exposed to antenatal steroids and exogenous surfactant and randomized to HFOV in the UKOS trial would have superior pulmonary function at follow-up to those ventilated conventionally. Some of the results of this study have been previously reported in the form of an abstract (18).
Study Population and Entry Criteria Twenty-five centers participated in the UKOS trial, including three outside of the UK. Infants were eligible for randomization if their gestational age was between 23 weeks and 28 weeks plus 6 days. Other eligibility criteria and the high frequency and CV strategies applied in the neonatal period are summarized in the online supplement and have been described in detail previously (12). The pulmonary function assessments at 1 year corrected age (age from expected date of delivery) were performed at a single center in London, UK, and a subgroup of trial infants was recruited from the participating centers that were within reasonable traveling distance from this center. Informed written consent from infants' parents was obtained before testing, and the study was approved by both the South Thames Multicenter Research Ethics Committee and the Local Research Ethics Committee of King's College Hospital National Health Service Trust.
Pulmonary Function Testing Protocol
Sample Size
Statistical Analysis
Subjects From the 12 centers that participated in this follow-up study, 185 infants were eligible for pulmonary function testing. From these, parents of 149 were invited to attend for testing. The remaining 36 infants were either living too far away from London or had been lost to follow-up. The parents of 90 infants agreed to participate in the follow-up study. However, 10 failed to attend their appointments, 3 (1 CV and 2 HFOV) were repeatedly unwell or remained dependent on supplemental oxygen, and 1 could not be successfully sedated. This left 76 infants who formed the study group. The studied infants had slightly lower mean birth weight and gestational age compared with the remainder of the trial survivors, as indicated by 95% confidence intervals that excluded zero but were otherwise similar with respect to a range of sociodemographic and clinical parameters (Table 1). Follow-up data were not available for all 592 survivors of the trial. The follow-up data in Table 1 were obtained exclusively from standardized respiratory questionnaires completed at 6 and 12 months corrected age by each infant's own pediatrician.
When split according to randomized mode of ventilation, the two pulmonary function groups were well matched for a range of baseline characteristics, with no statistically significant differences (Table 2). All follow-up data in Table 2 were obtained when each infant attended for pulmonary function testing.
Pulmonary Function Most infants had complete pulmonary function results. On some occasions, technically acceptable recordings were not obtained, or the infant woke before measurements were complete. Measurements of FRCpleth were missing for two infants (one in each group) and of FRCHe for four infants (one CV and three HFOV). One or other type of FRC measurement was available for all infants. Airway resistance measurements were missing for six infants (three in each group) and tidal breathing parameters for five infants (three CV and two HFOV). There were no statistically significant differences in pulmonary function between the two groups (Table 3). Moreover, the distribution of pulmonary function within both groups was comparable (Figures 1 and 2).
We have demonstrated that pulmonary function at follow-up did not differ significantly between infants who had been supported by HFOV and those who had been ventilated conventionally. These results are important, as they represent the only infant pulmonary function data published as follow-up in a randomized controlled trial of HFOV where lung volume optimization was employed. It is possible that differences between the groups might have emerged if other lung function tests, for example, the forced partial expiratory flow or the raised volume rapid thoracoabdominal compression maneuvers, had been included, as these techniques produce dynamic compression of the airways. We did not measure maximal flow at FRC, as at the time the UKOS trial was designed, we considered that the rapid thoracoabdominal compression technique produced results with unacceptably high intersubject and intrasubject variations. It is also possible that measurement at times when the infants had lower respiratory tract infection might have revealed differences between those who had received HFOV and those randomized to CV. The effects of infection on lung function, however, would have been very variable depending on the type of infecting organisms and the severity of the resultant illness, as well as any underlying abnormalities of lung function. As a consequence, a very large sample size would have been required to satisfactorily resolve this question. As this was a follow-up study with a finite-sized population, it was not possible to recruit additional infants when our sample size fell short of the planned 100. Despite a reduced sample size of 76, our results have shown that the maximum likely difference, deduced from the 95% confidence intervals, was low for four of the seven measurements: FRCpleth, 13%; FRCHe, 13%; FRCHe:FRCpleth, 6%; and respiratory rate, 18% (calculated from maximum confidence limit/mean measurement). We thought it was preferable that all pulmonary function testing was performed at a single center to ensure consistent results. It was therefore impractical to attempt to recruit all survivors from the UKOS trial, as infants had been enrolled in the trial from all over the UK, Ireland, Singapore, and Australia. We therefore recruited a subset of infants whose parents were prepared to travel to south London for the study. The subset was representative of the entire cohort of UKOS survivors, and when split according to randomized ventilator mode, the two groups were well matched for baseline characteristics. We suspected that parents might be more willing to participate in the follow-up study if their child had ongoing respiratory problems that might warrant pulmonary function testing, but this was not the case, as infants who were tested had similar rates of chronic lung disease and respiratory symptoms compared with those who were not tested. Previous studies have examined pulmonary function in infancy after treatment with HFOV in the neonatal period (16, 20, 21). A subset of infants who participated in the HIFI trial had pulmonary function measured at 9 months corrected age (16). As in our study, no difference was found in any aspect of pulmonary function between infants who had been randomized to HFOV and those randomized to CV. The HIFI trial has been criticized, however, for not employing a lung volume optimization strategy for infants on HFOV (15, 17), and the infants studied were very different from those in the UKOS trial, nearly all of whom received antenatal steroids and exogenous surfactant. A more recent study evaluated pulmonary function at 6 and 12 months corrected age in infants with chronic lung disease, some of whom had been exposed to HFOV in the neonatal period (21). At 12 months corrected age, infants who had received HFOV had a similar mean lung volume compared with those who had received CV but had significantly higher mean maximal expiratory flow at FRC. This apparent positive result must be treated with caution, as infants were allocated to mode of ventilation at the discretion of the attending clinician and depending on availability of equipment, rather than by randomization. Furthermore, only infants with chronic lung disease were studied, excluding a large proportion of very preterm neonates who would have required mechanical ventilation at birth but who did not subsequently develop chronic lung disease. Of the randomized controlled trials in premature infants comparing CV with elective HFOV using a high-volume strategy, only one, the Provo trial, has so far published detailed respiratory follow-up data, including pulmonary function assessments at between 5 and 8 years of age (22, 23). Unlike the UKOS trial, the Provo trial had found a difference in primary outcome in that more infants ventilated using HFOV had survived without chronic lung disease. There were, however, important differences between the two trials. All infants in the Provo trial received at least one dose of exogenous natural surfactant, but only 24% received antenatal steroids. Infants up to 36 weeks of gestation were eligible for the study and could be randomized to their allocated mode of ventilation as late as 12 hours after birth. Hence, this was a much less premature population of infants than that of the UKOS trial, and no attempt was made to minimize tidal breathing before initiation of HFOV. In the Provo respiratory follow-up study, which assessed 56% of all trial survivors, children who had been initially ventilated with CV had inferior pulmonary function compared with those initially ventilated with HFOV. These results implied that infants who were electively ventilated with HFOV had an improved respiratory outcome that persisted into childhood. Caution is warranted in interpreting these data, however, as infants receiving CV appeared to have an unusually poor outcome, evidenced by the fact that 49% of CV infants required supplemental oxygen at discharge, despite a relatively high mean birth weight of 1.46 kg. The aim of the UKOS pulmonary function follow-up study was to determine whether treatment with HFOV conferred any advantage with respect to infant pulmonary function, rather than to determine whether the pulmonary function of the studied infants was normal or abnormal. Infants in both groups showed a broad range of pulmonary function, with similar distributions of results between groups for all parameters measured. Because of the continuing improvements in infant pulmonary function techniques, previous published reference ranges may not be applicable to our study group, especially as they were generally based on data from healthy infants born at term (24, 25). There has furthermore been a trend toward lower normal values for FRCpleth over recent years (26). Therefore, although the mean lung volumes of the UKOS infants tested were close to those predicted by recently proposed reference equations (mean percentage predicted FRCpleth 99.4 and FRCHe 106.5) (22, 23), it is not possible to say with certainty whether these represent normal values. There are no recent published reference ranges available for inspiratory or expiratory airway resistance (25). In conclusion, we have found no significant difference in pulmonary function between preterm infants ventilated using high-frequency oscillation and those ventilated conventionally. These results and the lack of impact on chronic lung disease emphasize that prophylactic HFOV offers no advantage over CV with regard to respiratory outcome. Respiratory follow-up of all UKOS survivors at 2 years corrected age is currently underway.
The authors thank the infants and their parents for taking part in this study. They are indebted to the staff of all the UKOS participating centers and in particular to those from the centers listed here, which recruited infants for the pulmonary function study. They also thank Karl Sylvester, Andrew Theivendra, Annemarie van Overbeek, David Newby, and Metale Biswas for their help with the pulmonary function measurements. Recruiting centers (all in United Kingdom) include the following: Chelsea & Westminster Hospital, London; Guy's & St. Thomas' Hospital, London; King's College Hospital, London; Medway Maritime Hospital, Kent; Northwick Park Hospital, London; Nottingham City Hospital & Queen's Medical Centre, Nottingham; Princess Anne Hospital, Southampton; Queen Charlotte's Hospital, London; Rosie Maternity Hospital, Cambridge; St. George's Hospital, London; St. Peter's Hospital, Chertsey; and Southmead Hospital, Bristol.
Supported by the Medical Research Council, London, United Kingdom. This article has an online supplement, which is accessible from this issue's table of contents online at www.atsjournals.org Conflict of Interest Statement: M.R.T. has no declared conflict of interest; G.F.R. has no declared conflict of interest; E.S.L. has no declared conflict of interest; J.L.P. has no declared conflict of interest; S.A.C. has no declared conflict of interest; N.M. has been a co-recipient of a research grant from Serono Laboratories UK Ltd. and has been reimbursed by Abbot and by Chiesi for attending several conferences over the past five years and has received approximately ten lecture fees totaling £750 from Chiesi; A.D.M. has no declared conflict of interest; A.G. has received grants from SLE and Stephanie companies, and through national and international societies, and SLE has contributed to travel and hotel expenses when related to keynote lectures at national and international conferences. Received in original form October 19, 2003; accepted in final form December 20, 2003
This article has been cited by other articles:
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TOP
ABSTRACT
METHODS
