Pulmonary hemorrhage and exogenous surfactant therapy: A metaanalysis

• Published in: The Journal of pediatrics Vol. 123; no. 4; pp. 603 - 610
• Main Authors: Raju, Tonse N.K., Langenberg, Patricia
• Format: Journal Article
• Language: English
• Published: New York, NY Mosby, Inc 01.10.1993; Elsevier
• Subjects: Biological and medical sciences; Hemorrhage - chemically induced; Hemorrhage - epidemiology; Hemorrhage - etiology; Humans; Incidence; Infant, Newborn; Lung Diseases - chemically induced; Lung Diseases - epidemiology; Lung Diseases - etiology; Medical sciences; Pharmacology. Drug treatments; Pulmonary Surfactants - adverse effects; Pulmonary Surfactants - therapeutic use; Regression Analysis; Respiratory Distress Syndrome, Newborn - complications; Respiratory Distress Syndrome, Newborn - therapy; Respiratory system; Risk Factors; Human; Replacement therapy; Newborn; Pulmonary surfactant; Respiratory disease; Treatment efficiency; Respiratory distress; Hyaline membrane; Statistical study; Metaanalysis
• ISSN: 0022-3476; 1097-6833
• DOI: 10.1016/S0022-3476(05)80963-1

A metaanalysis of surfactant clinical trials was carried out to assess whether ornot an association exists between exogenous surfactant therapy and pulmonary hemorrhage. Trials that reported the pulmonary hemorrhage occurrence (group 1) and those that did not (group 2) were analyzed. Thirty-three treatment strategies were tested in 29 publications from 1980 through 1992. Eleven of these were group 1 trials, which reported a 3% overall incidence of pulmonary hemorrhage. The rates were significantly higher in both the treated and the control groups of natural surfactant trials than in synthetic surfactant trails (5.87% and 5.36% in the natural surfactant trials vs 2.51% and 1.04% in the synthetic surfactant trials, respectively). The pooled estimate of relative risk for pulmonary hemorrhage with any surfactant therapy was 1.47 (95% confidence interval 1.05, 2.07; p<0.05). Logistic regression modeling revealed that the nature of surfactant, treatment strategy, and lower mean birth weight had a significant influence on the relative risk of pulmonary hemorrhage; a similar trend was seen with higher mortality rates. Variation in the rates of patent ductus arteriosus did not have an independent effect on the estimated pulmonary hemorrhage risk. Most group 2 trials were published before 1990, and the median total sample size was 73, compared with 402 for the group 1 trials ( p<0.05), most of which were published in the 1990s. In 10 (50%) of 20 group 2 trials, pulmonary hemorrhage data were collected methodically, in comparison with all group 1 trials, most of which collected data prospectively. We conclude that pulmonary hemorrhage is a rare complication of respiratory distress syndrome. An awareness of the possible association of pulmonary hemorrhage with surfactant use in later trials and the differences in definitions and reporting practices probably explain variations in the reported incidence among the trials. The risk of pulmonary hemorrhage increases slightly, on an average of 47%, with any surfactant therapy. This increased risk is small compared with the documented benefits of surfactant therapy in respiratory distress syndrome.