Wes Onland, Moniek van de Loo, Martin Offringa, Anton van Kaam

Systematic reviews showed that systemic postnatal corticosteroids reduce the risk of bronchopulmonary dysplasia (BPD) in preterm infants. However, corticosteroids have also been associated with an increased risk of neurodevelopmental impairment. It is unknown whether these beneficial and adverse effects are modulated by differences in corticosteroid treatment regimens related to type of steroid, timing of treatment initiation, duration, pulse versus continuous delivery, and cumulative dose. To assess the effects of different corticosteroid treatment regimens on mortality, pulmonary morbidity, and neurodevelopmental outcome in very low birth weight infants. We conducted searches in September 2022 of MEDLINE, the Cochrane Library, Embase, and two trial registries, without date, language or publication- type limits. Other search methods included checking the reference lists of included studies for randomized controlled trials (RCTs) and quasi-randomized trials. We included RCTs comparing two or more different treatment regimens of systemic postnatal corticosteroids in preterm infants at risk for BPD, as defined by the original trialists. The following comparisons of intervention were eligible: alternative corticosteroid (e.g. hydrocortisone) versus another corticosteroid (e.g. dexamethasone); lower (experimental arm) versus higher dosage (control arm); later (experimental arm) versus earlier (control arm) initiation of therapy; a pulse-dosage (experimental arm) versus continuous-dosage regimen (control arm); and individually-tailored regimens (experimental arm) based on the pulmonary response versus a standardized (predetermined administered to every infant) regimen (control arm). We excluded placebo-controlled and inhalation corticosteroid studies. Two authors independently assessed eligibility and risk of bias of trials, and extracted data on study design, participant characteristics and the relevant outcomes. We asked the original investigators to verify if data extraction was correct and, if possible, to provide any missing data. We assessed the following primary outcome: the composite outcome mortality or BPD at 36 weeks' postmenstrual age (PMA). Secondary outcomes were: the components of the composite outcome; in-hospital morbidities and pulmonary outcomes, and long-term neurodevelopmental sequelae. We analyzed data using Review Manager 5 and used the GRADE approach to assess the certainty of the evidence. We included 16 studies in this review; of these, 15 were included in the quantitative synthesis. Two trials investigated multiple regimens, and were therefore included in more than one comparison. Only RCTs investigating dexamethasone were identified. Eight studies enrolling a total of 306 participants investigated the cumulative dosage administered; these trials were categorized according to the cumulative dosage investigated, 'low' being < 2 mg/kg, 'moderate' being between 2 and 4 mg/kg, and 'high' > 4 mg/kg; three studies contrasted a high versus a moderate cumulative dose, and five studies a moderate versus a low cumulative dexamethasone dose. We graded the certainty of the evidence low to very low because of the small number of events, and the risk of selection, attrition and reporting bias. Overall analysis of the studies investigating a higher dose versus a lower dosage regimen showed no differences in the outcomes BPD, the composite outcome death or BPD at 36 weeks' PMA, or abnormal neurodevelopmental outcome in survivors assessed. Although there was no evidence of a subgroup difference for the higher versus lower dosage regimens comparisons (Chi2 = 2.91, df = 1 (P = 0.09), I2 = 65.7%), a larger effect was seen in the subgroup analysis of moderate-dosage regimens versus high-dosage regimens for the outcome cerebral palsy in survivors. In this subgroup analysis, there was an increased risk of cerebral palsy (RR 6.85, 95% CI 1.29 to 36.36; RD 0.23, 95% CI 0.08 to 0.37; P = 0.02; I² = 0%; NNTH 5, 95% CI 2.6 to 12.7; 2 studies, 74 infants). There was evidence of subgroup differences for higher versus lower dosage regimens comparisons for the combined outcomes death or cerebral palsy, and death and abnormal neurodevelopmental outcomes (Chi2 = 4.25, df = 1 (P = 0.04), I2 = 76.5%; and Chi2 = 7.11, df = 1 (P = 0.008), I2 = 85.9%, respectively). In the subgroup analysis comparing a high dosage regimen of dexamethasone versus a moderate cumulative-dosage regimen, there was an increased risk of death or cerebral palsy (RR 3.20, 95% CI 1.35 to 7.58; RD 0.25, 95% CI 0.09 to 0.41; P = 0.002; I² = 0%; NNTH 5, 95% CI 2.4 to 13.6; 2 studies, 84 infants; moderate-certainty evidence), and death or abnormal neurodevelopmental outcome (RR 3.41, 95% CI 1.44 to 8.07; RD 0.28, 95% CI 0.11 to 0.44; P = 0.0009; I² = 0%; NNTH 4, 95% CI 2.2 to 10.4; 2 studies, 84 infants; moderate-certainty evidence). There were no differences in outcomes between a moderate- and a low-dosage regimen. Five studies enrolling 797 infants investigated early initiation of dexamethasone therapy versus a moderately early or delayed initiation, and showed no significant differences in the overall analyses for the primary outcomes. The two RCTs investigating a continuous versus a pulse dexamethasone regimen showed an increased risk of the combined outcome death or BPD when using the pulse therapy. Finally, three trials investigating a standard regimen versus a participant-individualized course of dexamethasone showed no difference in the primary outcome and long-term neurodevelopmental outcomes. We assessed the GRADE certainty of evidence for all comparisons discussed above as moderate to very low, because the validity of all comparisons is hampered by unclear or high risk of bias, small samples of randomized infants, heterogeneity in study population and design, non-protocolized use of 'rescue' corticosteroids and lack of long-term neurodevelopmental data in most studies. The evidence is very uncertain about the effects of different corticosteroid regimens on the outcomes mortality, pulmonary morbidity, and long term neurodevelopmental impairment. Despite the fact that the studies investigating higher versus lower dosage regimens showed that higher-dosage regimens may reduce the incidence of death or neurodevelopmental impairment, we cannot conclude what the optimal type, dosage, or timing of initiation is for the prevention of BPD in preterm infants, based on current level of evidence. Further high quality trials would be needed to establish the optimal systemic postnatal corticosteroid dosage regimen.