Background - Methods - Results - Characteristics of Included Studies - References - Data Tables & Graphs
Lung disease is a major cause of death in very low birth weight babies. Chronic lung disease (CLD) following mechanical ventilation for babies with breathing difficulties is also common. It is possible that the low gas exchange in newborns during breathing may help cause chronic lung disease. Elective high frequency jet ventilation (HFJV) is one type of mechanically assisted breathing method that may improve gas exchange in neonates without injuring the lung. The review of trials found there may be benefits of HFJV but not enough evidence of adverse effects. More research is needed.
In the overall analysis, there were no significant differences in the incidence of neonatal mortality, IVH all grades or in grades 3 or 4 IVH. In the subgroup where 'low volume strategy' was used there was a non-significant trend toward an increase in risk of IVH all grades and grades 3 or 4 IVH.
There is evidence from animal studies (Barringer 1982, Carlon 1983, Hoff 1981) and adult studies (Carlon 1981 and Turnbull 1981) that high frequency jet ventilation (HFJV, 200 - 400 breaths per minute) may reduce the severity of lung injury during mechanical ventilation. Observational studies (Pokora 1983, Carlo 1984) suggested that HFJV may improve gas exchange in neonates with respiratory failure with lower pressures than conventional ventilation (CV).
The following 'a priori' subgroup analyses were planned:
1) Trials with and without surfactant replacement therapy. Surfactant replacement therapy would be expected to increase alveolar recruitment, attentuate RDS and thus could alter treatment effects.
2) High and low volume ventilator strategies on HFJV. Ventilator strategies aimed at maintaining effective lung volume (high volume strategy, HVS) such as use of high mean airway pressures, manoeuvres to recruit alveolar volume after suctioning and weaning of FiO2 before pressure would be expected to have better responses to treatment than those using a low volume strategy (LVS).
3) Infants at different gestational ages and birth weight, as they have different baseline rates of CLD and neurologic injury, may respond differently to the treatment.
4) Trials with and without adequate humidification. Different rates of necrotizing tracheitis and its consequences would be expected due to treatment.
1) Mortality by 28-30 days, and before discharge.
2) Chronic lung disease defined as; a) supplemental oxygen or MV at
28d; or b) supplemental oxygen or MV at 36 weeks post-menstrual age (PMA)
3) Pulmonary air leak syndromes defined as presence of any of the following:
pneumothorax, pulmonary interstitial emphysema or pneumoperitonium.
4) Intraventricular hemorrhages; a) all grades; b) grades 3 or 4
5) Periventricular leukomalacia
6) Periventricular echodensities
7) Necrotizing tracheobronchitis or subglottic stenosis
8) Pulmonary and neurodevelopmental outcomes in childhood
Additional data were obtained from Keszler 1997.
Methods used to collect data from the included trials:
Each author extracted data separately, then compared and resolved differences.
Additional data were requested from authors as required.
Methods used to synthesize the data:
The standard method of the Neonatal Review Group was used, including
for categorical data, use of relative risk (RR) and risk difference (RD).
From 1/RD the number needed to treat (NNT) for benefits or number needed
to harm (NNH) for adverse effects and their 95% CI's were calculated. For
continuous data, standardized mean difference with 95% CI was used.
In two trials the intention was to use a lower mean airway pressure (LVS) when switching from CV to HFJV (Carlo 1990, Wiswell 1996) while the third trial (Keszler 1997) the intention was to use a higher mean airway pressure (HVS).
Gestational age is normally calculated as postmenstrual age (PMA) at birth and together with postnatal age, is used here to define O2 dependency at 36 weeks. Wiswell 1996 and Keszler 1997 used the term postconceptual age but did not define it. This needs author clarification.
Surfactant was administered to infants in two trials (Wiswell 1996, Keszler 1997).
Details of each study are given in the Included Studies Table and in the references.
CLD at 36 weeks PMA is significantly reduced in the two trials that assessed this outcome (Wiswell 1996 and Keszler 1997) both of which were also in the surfactant subgroup [summary RR 0.58 (0.34, 0.98); RD 0.138 (-0.268, -0.007), NNT 7 (4, 90)]. This outcome is also reduced in the trial (Keszler 1997) where high volume strategy was used [RR 0.50 (0.27, 0.92); RD -0.204 (-0.374, -0.034), NNT 7 (4, 90)]. The one trial (Wiswell 1996) that used a low volume strategy when on HFJV and reported CLD at 36 weeks PMA found no difference.
The use of home oxygen was assessed in only one study (Keszler 1997). The number of patients on home oxygen was found to be lower in the HFJV group [RR 0.24 (0.07, 0.79); RD -0.176 (-0.306, -0.047), NNT 5 (3, 21)].
Air Leak Syndromes
There is no significant difference in incidence of air leak syndromes
in the individual trials or in the overall analysis.
Neonatal Mortality
There is no significant difference in neonatal mortality in any individual
trial, in the overall analysis, or in the subgroup analyses.
Intraventricular Hemorrhage (IVH)
There are no significant differences in the incidence of IVH of all
grades in any individual trial or in the overall analysis. In the subgroup
where low volume strategy was used there is a trend towards an increase
in incidence of IVH. There are no significant differences in incidence
of the more severe grades of IVH (3 or 4) in any individual trials or in
the overall analyses [summary RR 1.37 ( 0.79, 2.37)]. In the subgroup where
a low volume strategy was used only one trial reported IVH by grade (Wiswell
1996) and there is a non-significant trend towards an increase in grades
3 or 4 IVH. In the trial by Carlo 1990 progression
to IVH grades 2 to 4 was reported as 9/21 in HFJV group and 7/21 in the
CV group. However these figures include two patients in the HFJV group
and seven patients in the CV group without initial head ultrasound who
also developed grade 2 to 4 IVH. Thus these results were not included in
the meta-analysis.
Periventricular Leukomalacia (PVL)
This outcome was reported in two studies (Wiswell
1996, Keszler 1997). Overall there is a
non-significant trend towards an increase in risk [summary RR 1.24 (0.59,
2.61)] of PVL. In the one trial where a high volume strategy was intended
(Keszler 1997) there is a trend towards a reduction
in risk of PVL [RR 0.42 (0.14, 1.30)]. In the subgroup where a low volume
strategy was used, only one trial (Wiswell 1996)
reported PVL and there is a significant increase [RR 5.0 (1.19, 21.04),
RD 0.250 (0.069, 0.0431), NNH 4.0 (2.3,14.5)].
Periventricular Echodensities (PVE)
This outcome was assessed only in the Wiswell
1996 trial and no significant difference was found.
Continuous Data Outcomes
The following continuous outcomes were reported from the eligible trials
and have been included after reviewing the studies. Keszler
1997 also reported medians and ranges for the continuous data since
the data were not normally distributed.
Days in supplemental oxygen was assessed by all the trials as means and SD. The pooled results show a non-significant trend towards a decrease with HFJV. Keszler 1997 also reported the medians and ranges of days in supplemental oxygen [HFJV 37 (3-160) and CV 46 (3-167)].
Days on Mechanical ventilation (MV) is available from all three trials as means and SD. The pooled data gave a non-significant trend towards an increase in the days on MV. Keszler 1997 also reported the data as medians and ranges [HFJV 20 (3-96) and CV 26 (2-82)].
Wiswell 1996 reported a non-significant trend towards a decrease in length of hospital stay. Keszler 1997 reported the medians and ranges as 71 (24-198) for the HFJV group and 76 (33-167) for the CV group, which were not significantly different.
Two of the planned subgroup analyses could not be performed. There were no outcome data presented by gestational age or birth weight despite stratification at randomization in two of the studies (Wiswell 1996, Keszler 1997). Lack of humidification is now thought to be only of historical importance and was not a reported problem in any of the included trials.
Pulmonary Outcomes
Overall this review suggests that the elective use of HFJV for preterm
infants with RDS is associated with a reduction in measures of CLD such
as oxygen or ventilator dependency at 36 weeks PMA and use of oxygen therapy
at home. However, there are no long term pulmonary follow-up data from
these studies. In the excluded trial (Carlo 1987)
there were no differences in the incidence of CLD between the groups. During
the 48 hour study period, 4/20 in the HFJV group and 8/20 in the CV developed
one or more air leaks, this was, however, not significant.
Neurological Outcomes
Although there was no significant difference in short term brain injury
in the overall analysis, subgroup analysis indicated that brain injury
was more common when a 'low volume strategy' for HFJV was used (Wiswell
1996). The Keszler 1997 trial intended
to use a 'high volume strategy' but this was actually used in only 56%
of infants. Post hoc analyses indicated that the rate of acute brain injury
such as grades 3 or 4 IVH was higher in the group who actually received
a 'low volume strategy' (high volume strategy 3/34, 8.8% vs low volume
strategy 6/27, 22.2%). There were no differences in the incidence of IVH
developing during the 48 hour study period in the trial by Carlo
1987.
In the overall analysis there was a trend towards an increase in the incidence of PVL which was not significant. However, in the trial by Wiswell 1996 there was a significant increase in the incidence of PVL. In the report of this trial the logistic regression analysis of ventilator assignment showed that HFJV was independently associated with PVL. Hypocarbia was found not to be independently associated with any adverse outcomes in this trial although observational studies have suggested that marked hypocarbia is associated with an increased risk of cystic PVL (Calvert 1986, Graziani 1992, Fujimoto 1994).
The cause of acute neurological injury in some infants on HFJV is uncertain. This outcome has been reported from a single centre and the generalizability of this result is uncertain. The possible association with the use of a 'low volume strategy' during high frequency ventilation here has also been suggested in the review of trials evaluating high frequency oscillatory ventilation (Henderson-Smart 2002). The reason for this apparent association is unknown. To resolve the issue a randomized controlled trial comparing high and low volume strategies would be required. This is unlikely to be done given the current preference to use the 'high volume strategy', a strategy which has been recommended on the basis of animal studies (Froese 1991).
Benefits in terms of a reduction in CLD, where on an average seven infants would be needed to be treated to prevent CLD at 36 weeks in one infant, may be associated with increased risk of PVL, where for every four infants treated there would be one additional infant with PVL. This adverse effect is a 'worst case' scenario as it is based on one study (Wiswell 1996), and the other study (Keszler 1997) reporting this outcome showed a trend toward a reduction in PVL. Until these differences can be clarified it remains possible that the risks of elective HFJV may outweigh the benefits.
It is of concern that there were no long term neurodevelopmental outcomes available from any of these trials.
| Study | Methods | Participants | Interventions | Outcomes | Notes | Allocation concealment |
| Carlo 1990 | Single centre
Concealment at randomization - Yes Blinding of intervention - No Complete followup - Yes ( 93%) Blinding of outcome - Can't tell |
Forty-five preterm infants less than 24 hours of age with RDS and stratified into 3 groups; 1000-1250g, 1251-1500g, 1501-2000g | HFJV, PIP and Paw decreased by 20% when changed to HFJV - low volume
strategy (LVS), 250/min 1:3 IT:ET ratio; backup use of alternative intervention
in both directions was permitted if failure documented on two consecutive
ABG.
Controls were ventilated with CV on time cycled pressure limited ventilators (Bear Cub). The protocol for CV was not specified. |
Mortality at 28d, CLD at 28d, ALS, progression of IVH, success after crossover, days on MV, days on supplemental oxygen, MV at 28d. | Mean age at randomization 15.5 hrs for HFJV and 14 hrs for CV; enrollment stopped when there was no difference, no surfactant used. | A |
| Keszler 1997 | Multicentre trial; 8 centres
Blinding of randomization -Yes Blinding of intervention - No Completeness of followup - Yes (91%); 14 patients from one centre were excluded from the analysis as they were simultaneously enrolled in another similar trial and reported in the study by Wiswell 1996. Of 130 patients analysed, 61(94%) were assessed for IVH and PVL in the HFJV group and 59(91%) in CV group. Blinding of outcome measures -Yes |
Reported on 130 of 144 preterm infants stratified into 3 groups 700-1500g, 1001-1250g, 1251-1500g; <36weeks gestational age; <20 hours of age and MV for <12 hours of age; | HFJV Bunnell Life Pulse; concurrent CV used for humidification, positive
end expiratory pressure (PEEP) for alveolar recruitment and to provide
intermittent sigh breaths in form of background intermittent mandatory
ventilation; at the time of randomization PEEP increased and this increased
Paw 0.5 to 2 cms H2O, (HVS).
CV was at rates between 30-60/min, inspiratory time of 0.3 to 0.4 sec and PEEP of 4 to 6 cms of water. |
Mortality; CLD at 28d and 36 w PCA, O2 at discharge, days in O2, IVH all grades and grades 3 or 4, PVL, Days in hospital, crossover success | Surfactant replacement therapy used. Although high volume strategy was the intended method, some centres used a low volume strategy (44% of infants). Concerns about adverse effects in some patients in another trial forced early stoppage of the trial. Mean age at randomisation 8.3+4.2 hrs, antenatal corticosteroids usage 30% CV vs 21% HFJV | A |
| Wiswell 1996 | Single centre
Concealment of randomization -Yes Blinding of intervention - No Completeness of followup -Yes (100%) Blinding of outcome assessment -Yes |
Seventy-three infants less than 33 weeks GA, > 500 gms, < 2000 gms; < 24 hours of age; with RDS | HFJV ( Bunnell Life Pulse) 420/min; PIP decreased to 80-90% of CV (LVS) when switched over, PEEP unchanged. Protocol for CV was not specified. | IVH grades 3 or 4; IVH all grades; periventricular echodensities (PVE); cystic periventricular leukomalacia (CPVL); supplemental oxygen at 28d and 36 wk; mortality at 28d and 36wk; days on MV and days in hospital. | Surfactant therapy used; backup use of the alternative intervention was allowed in both directions. Mean age at randomization 7.2+6 hrs., antenatal corticosteroids HFJV 22% CV 19%. | A |
| Study | Reason for exclusion |
| Carlo 1987 | Intervention commenced early but was only for 48 hours and then all the patients were switched back to conventional ventilation. The trial is therefore not comparable to trials where the intervention was intended to be used until extubation or failure to ventilate adequately. Infants with birth weights < 1000 gms, who are of most interest in this review because of their high rate of CLD, IVH and PVL, were excluded from the study. |
| Keszler 1991 | Patients who developed pulmonary interstitial emphysema on conventional ventilation were randomized. This trial thus fulfilled the definition of rescue therapy and is the subject of another review. |
Carlo WA, Siner B, Chatburn RL, Robertson S, Martin RJ. Early randomized intervention with high-frequency jet ventilation in respiratory distress syndrome. J Pediatr 1990;117:765-70.
Keszler 1997 {published data only}
Keszler M, Modanlou HD, Brudno DS, Clark FI, Cohen RS, Ryan RM, Kaneta MK, Davis JM. Multicenter controlled clinical trial of high frequency jet ventilation in preterm infants with uncomplicated respiratory distress syndrome. Pediatrics 1997;(4):593-597.
Wiswell 1996 {published data only}
Wiswell TE, Graziani LJ, Kornhauser MS, Cullen J, Merton DA, McKee L, Spitzer AR. High-Frequency jet ventilation in the early management of respiratory distress syndrome is associated with a greater risk for adverse outcomes. Pediatrics 1996;98:1035-1043.
Carlo WA, Chatburn RL, Martin RJ. Randomised trial of high frequency jet ventilation versus conventional ventilation in respiratory distress syndrome. J Pediatr 1987;110:275-82.
Keszler 1991 {published data only}
Keszler M, Donn SM, Bucciarelli RL, et al. Multicentre controlled trial comparing high-frequency jet ventilation and conventional ventilation in newborn infants with pulmonary interstitial emphysema. J Pediatr 1991;119:85-93.
* indicates the primary reference for the study
Barringer M, Meredith J, Prough D et al: Effectiveness of high frequency jet ventilation in management of experimental bronchopulmonary fistula. Am J Surg 1982;48:610-613.
Calvert SA, Hoskins EM, Fong KW, et al. Etiological factors associated with the development of periventricular leukomalacia. Acta Paediatr Scand 1986;76:254-259.
Carlo WA, Chatburn RL, Martin RJ et al. Decrease in airway pressure during high-frequency jet ventilation in infants with respiratory distress syndrome. J Pediatr 1984;101-107.
Carlon G, Kahn R, Howland W et al: Clinical experience with high frequency jet ventilation. Crit Care Med 1981;9(1):1-6.
Carlon GC, Griffin J, Cole R, et al: High frequency jet ventilation in experimental airway disruption. Crit Care Med 1983;11:353-355.
Ehrenkranz RA, Mercurio MR. Bronchopulmonary dysplasia. In: Sinclair JC, Bracken MB, editor(s). Effective care of the newborn infant. Oxford: Oxford University Press, 1992.
Froese AB, Bryan AC. Reflections on the HIFI trial. Pediatrics 1991;87:565-567.
Fujimoto S, Togari H, Yamaguchi N et al. Hypocarbia and cystic periventricular leukomalacia in premature infants. Arch Dis Child 1994;71:F107-F110.
Graziani LJ, Spitzer AR, Mitchell DG et al. Mechanical ventilation in preterm infants: neurosonographic and developmental studies. Pediatrics 1992;90:515-522.
Henderson-Smart DJ, Bhuta T, Cools F, Offringa M. Elective high frequency oscillatory ventilation vs conventional ventilation in preterm infants with acute pulmonary dysfunction (Cochrane Review). In: The Cochrane Library, Issue 3, 2002. Oxford: Update Software.
Hoff B, Smith R, Wilson E et al. High frequency ventilation during bronchopleural fistula. Anesthesiology 1981;55:A71.
Jobe AH, Ikegami M. Lung development and function in preterm infants in the surfactant treatment era. Ann Rev Physiol 2000;62:825-846.
Kuban KCK, Leviton A. Cerebral palsy. N Engl J Med 1994;330:188-195.
Pokora T, Bing D, Mammel M, Boros S. Neonatal high frequency jet ventilation. Pediatrics 1983;72:27-32.
Turnbull A, Carlon G, Howland W et al. High frequency jet ventilation in major airway or pulmonary disruption. Ann Thorac Surg 1981;32(5):468-74.
Bhuta T, Henderson-Smart DJ. Elective high frequency jet ventilation versus conventional ventilation for respiratory distress syndrome in preterm infants (Cochrane Review). In: The Cochrane Library, Issue 2, 1998.
02 Elective HFJV vs CV with surfactant replacement therapy
02.01 CLD at 28 d in survivors
02.02 CLD at 36 w in survivors
02.03 Air leak syndromes
02.04 Home oxygen in survivors
02.05 Neonatal mortality
02.06 IVH all grades
02.07 IVH grades 3 or 4
02.08 Periventricular echodensities
02.09 Periventricular leukomalacia
02.10 Days in oxygen
02.11 Days in hospital
02.12 Days on IPPV
03 Elective HFJV with low volume strategy
03.01 CLD at 28 d in survivors
03.02 CLD at 36 w in survivors
03.03 Air leak syndromes
03.04 Home oxygen in survivors
03.05 Neonatal mortality
03.06 IVH all grades
03.07 IVH grades 3 or 4
03.08 Periventricular echodensities
03.09 Periventricular leukomalacia
03.10 Days in oxygen
03.11 Days in hospital
03.12 Days on IPPV