Abstract

A total of 346 infants aged less than 72 hours were randomly allocated to be treated either by high frequency positive pressure ventilation (HFPPV; rate fixed at 60/minute throughout treatment and initial inspiratory:expiratory (I:E) ratio 1:2, increased to 1:1 if necessary) or by low frequency positive pressure ventilation (LFPPV; rate less than or equal to 40/minute and initial I:E ratio usually 1:1, both decreasing during weaning). The main hypotheses were that HFPPV reduces pneumothorax, chronic lung disease and death before discharge in all infants, as well as those with hyaline membrane disease, and that it reduces the incidence of later neurodevelopmental complications in infants of less than 33 weeks' gestation. Among all the infants the rate of pneumothorax was 19% in the HFPPV group and 26% in the LFPPV group (p = 0.13; odds ratio 0.7, 95% confidence intervals (CI) 0.4 to 1.1); there was no difference in mortality or the incidence of chronic lung disease. In infants of less than 33 weeks' gestation there were no differences in adverse neurodevelopmental outcomes. Among the subgroup of 237 infants with hyaline membrane disease, median fractional inspired oxygen at the time of entry to the trial was 0.6 in the HFPPV group and 0.7 in the LFPPV group, indicating that many had moderately severe disease. In patients with hyaline membrane disease HFPPV was associated with a lower rate of pneumothorax (18% in the HFPPV group compared with 33% in the LFPPV group, p = 0.013, odds ratio 0.5, 95% CI 0.3 to 0.8, with no differences in mortality, or in duration of intubation or supplementary oxygen in survivors. As used in this study, HFPPV was the preferred ventilator regimen for infants with hyaline membrane disease.