The adverse neuro-developmental effects of postnatal steroids in the preterm infant: a systematic review of RCTs

Abstract

Background: Recent reports have raised concerns that postnatal steroids may cause neuro-developmental impairment in preterm infants. This systematic review was performed with the objective of determining whether glucocorticoid therapy, to prevent or treat bronchopulmonary dysplasia, impairs neuro-developmental outcomes in preterm infants.

Method: A systematic review of the literature was performed. Medline was searched and articles retrieved using predefined criteria. Data from randomized controlled trials with adequate neuro-developmental follow up (to at least one year) were entered into a meta-analysis to determine the effects of postnatal treatment of preterm infants with glucocorticoids. Cerebral palsy rates, and neuro-developmental impairment (developmental score more than 2SD below the mean, or cerebral palsy or blindness) were analyzed. The studies were divided into 2 groups according to the extent of contamination of the results by treatment of controls with steroids after the initial study period, those with less than 30% contamination, and those with more than 30% contamination or size of contamination not reported.

Results: Postnatal steroid therapy is associated with an increase in cerebral palsy and neuro-developmental impairment. The studies with less contamination show a greater effect of the steroids, consistent with a real direct toxic effect of steroids on the developing central nervous system. The typical relative risk for the development of cerebral palsy derived from studies with less than 30% contamination is 2.86 (95% CI 1.95, 4.19). The typical relative risk for the development of neuro-developmental disability among followed up infants from studies with less than 30% contamination is 1.66 (95% CI 1.26, 2.19). From this subgroup of studies, the number of premature infants who need to be treated to have one more infant with cerebral palsy (number needed to harm, NNH) is 7; to have one more infant with neuro-developmental impairment the NNH is 11.

Conclusions: Postnatal pharmacologic steroid treatment for prevention or treatment of bronchopulmonary dysplasia is associated with dramatic increases in neuro-developmental impairment. As there is no clear evidence in the literature of long term benefit, their use for this indication should be abandoned.

Figure 1

Meta-analysis of the effects of postnatal steroids on death. Studies are displayed in order of the degree of known contamination of the randomization, Fitzhardinge and O'Shea known to have no contamination, Yeh, Shinwell, and Jones with progressively increasing degrees of contamination, and Cummings, Subhedar and Vincer unknown.

Figure 2

Meta-analysis of the effects of postnatal steroids on neurodevelopmental impairment, among all randomized infants. Studies are displayed in order of the degree of known contamination of the randomization, Fitzhardinge and O'Shea known to have no contamination, Yeh, Shinwell, and Jones with progressively increasing degrees of contamination, and Cummings, Subhedar and Vincer unknown.

Figure 3

Meta-analysis of the effects of postnatal steroids on neurodevelopmental impairment among surviving, followed up, infants. Studies are displayed in order of the degree of known contamination of the randomization, Fitzhardinge and O'Shea known to have no contamination, Yeh, Shinwell, and Jones with progressively increasing degrees of contamination, and Cummings, Subhedar and Vincer unknown.

Figure 4

Meta-analysis of the effects of postnatal steroids on cerebral palsy, among all randomized infants. Studies are displayed in order of the degree of known contamination of the randomization, Fitzhardinge and O'Shea known to have no contamination, Yeh, Shinwell, and Jones with progressively increasing degrees of contamination, and Cummings, Subhedar and Vincer unknown.

Figure 5

Meta-analysis of the effects of postnatal steroids on cerebral palsy among surviving, followed up, infants. Studies are displayed in order of the degree of known contamination of the randomization, Fitzhardinge and O'Shea known to have no contamination, Yeh, Shinwell, and Jones with progressively increasing degrees of contamination, and Cummings, Subhedar and Vincer unknown.