a trial from 26 to 29+6 treatment wit weeks(RR 0.49, 95%CI 0.34 to 0.72, two studies, 242 infants), changes in the rates of maternal death, maternal infection, fetal 30 to 32 +6 weeks(RR 0.56, 95% CI 0.36 to 0.87, two studies, death, neonatal chronic lung disease or birthweight. Treatment 361 infants)and 33 to 34 +6 weeks(RR 0.53, 95%Cl 0.31 to with antenatal corticosteroids is also associated with a reduction 0.91, two studies, 434 infants), but not from less than 26 weeks in the incidence of neonatal necrotising enterocolitis and systemic RR 2.86, 95% CI 0.37 to 21.87, one study, 24 infants), 35 to infections in the first 48 hours of life, as well as a reduction in 36+ 6 weeks(RR 0.61, 95%CI O11 to 3. 26, one study, 189 the need for respiratory support or neonatal intensive care unit infants)and less than 36 weeks. Cerebroventricular haemorrhage admission. However, one trial (Amorim 1999)recruiting women as significantly reduced in corticosteroid-treated infants enter- with severe preeclampsia, using a protocol that included repeat ing a trial from 26 to 29+6 weeks(RR 0.45, 95%CI 0.21 to weekly courses of antenatal betamethasone if the women remained 0.95, one study, 227 infants), but not from less than 26 weeks(RR undelivered, suggested that the treated women were at increased 1. 20,95%CI 0. 24 to 6.06, one study, 27 infants), 30 to 32+6 risk of gestational diabetes. The women in this trial had a fasting weeks(RR 0.23, 95% CI 0.03 to 2.00, one study, 295 infants), glucose tolerance test more than 72 hours after the initiation of 33 to 34+6 weeks(RR 1.11, 95% CI 0. 23 to 5.40, one study, the study treatment if they were undelivered; 123(56%)women 339 infants), 35 to 36+6 weeks(no events in 191 infants)and under went the glucose tolerance test. It may not be appropriate to greater than 36 weeks(no events in 42 infants). Birthweight was generalise this to women without pre-eclampsia. It is also difficult significantly decreased in infants entering a trial from 30 to 32 to determine whether the fact that the protocol in this study used 6 weeks(FWMD-19064 grams, 95%CI-35998 to-2130 weekly repeat courses of antenatal corticosteroids was of relevance grams, one study, 319 infants), but not from less than 26 weeks to the outcome ms,95%CI-607 37 to 733.65 grams,on study, 49 infants), 26 to 29+6 weeks(FWMD 264l grams, 95% Concern has been expressed as to whether antenatal corticosteroids Cl-215.55 to 268.37 grams, one study, 261 infants ) 33 to 34+ are beneficial in the current era of advanced neonatal practice, 6 weeks(FWMD-3872 grams, 95%CI-17229 to 94.85 grams, conclusions concerning their benefits one study, 353 infants), 35 to 36+6weeks(FWMD-1357 grams, were based mainly on data from the 1970s. This update shows 95% CI.45 to 148.31 grams, one study, 194 infants) and that combined fetal and neonatal death, neonatal death, RDS and cerebroventricular h 418.67 grams, one study, 42 infants ). No statistically significant subgroup of trials conducted in the 1990s. These trials contributed differences between groups treated with antenatal corticosteroids 26% of the overall data to the review. This supports the continued and controls were seen for combined fetal and neonatal deaths or use of antenatal corticosteroids fetal deaths alone in the different subgroups of gestational age at The gestational age range at which antenatal corticosteroids pro- rial entry examined. vide benefit has been subject to debate, with some reviews sug- Antenatal corticosteroids versus placebo or no treatment(by gesting no benefit at less than 28 weeks(Crowley 1996). Previ- presence or absence in protocol of weekly repeat doses of corti- ously the effect of antenatal corticosteroids has been examined by oster sed on gesta rational delivery. Th Data were available by the presence or absence in the protocol that antenatal corticosteroids reduce the incidence of cerebroven- of weekly repeat doses of corticosteroid if the mother remained tricular haemorrhage even in those infants born before 28 weeks es that relate to the However, this review also examined outcomes by subgroups based mother and fetus/neonate. There was no difference ineffect of cor- on the clinically more relevant measure; gestational age at first ticosteroid treatment on chorioamnionitis, puerperal sepsis, com- dose of treatment(gestational age at trial entry). RDS is reduced bined fetal and neonatal death, fetal death, neonatal death. rds when corticosteroids are first given at 26 to 29.9 weeks, 30 to 32.9 or cerebroventricular haemorrhage between studies which used a weeks and 33 to 34.9 weeks. Furthermore, both cerebroventricular of antenatal corticosteroid and studies that allowed haemorrhage and neonatal death are reduced at 26 to 29. 9 weeks eekly repeats if the wor emained undelivered No difference is shown for primary outcomes at gestational ages of less than 26 weeks. While eight trials included in this review recruited pregna from less than 26 weeks gestation, and a DISCUSSI。N further three did not specify the lower gestational age for entry, only one trial (n=49 infants)contributed data to this review at The results of the 21 studies included in this updated review cat- this extreme gestation. egorically support the conclusion of the previous review(Crow- Antenatal corticosteroid use reduces neonatal death even when ley 1996), that treatment with antenatal corticosteroids reduces infants are born less than 24 hours after the first dose has been neonatal death, respiratory distress syndrome(RDS), and cere- given Reduction in RDS is seen in infants born up to seven days broventricular haemorrhage in preterm infants. Furthermore, after the first dose. This review has not shown any benefit in pri- Antenatal corticosteroids for accelerating fetal lung maturation for women at risk of preterm birth( Review) Copyright @2006 The Cochrane Collaboration. Published by John wiley& Sons, Ltdin corticosteroid-treated infants entering a trial from 26 to 29 + 6 weeks (RR 0.49, 95% CI 0.34 to 0.72, two studies, 242 infants), 30 to 32 + 6 weeks (RR 0.56, 95% CI 0.36 to 0.87, two studies, 361 infants) and 33 to 34 + 6 weeks (RR 0.53, 95% CI 0.31 to 0.91, two studies, 434 infants), but not from less than 26 weeks (RR 2.86, 95% CI 0.37 to 21.87, one study, 24 infants), 35 to 36 + 6 weeks (RR 0.61, 95% CI 0.11 to 3.26, one study, 189 infants) and less than 36 weeks. Cerebroventricular haemorrhage was significantly reduced in corticosteroid-treated infants enter￾ing a trial from 26 to 29 + 6 weeks (RR 0.45, 95% CI 0.21 to 0.95, onestudy, 227 infants), but not from less than 26 weeks (RR 1.20, 95% CI 0.24 to 6.06, one study, 27 infants), 30 to 32 + 6 weeks (RR 0.23, 95% CI 0.03 to 2.00, one study, 295 infants), 33 to 34 + 6 weeks (RR 1.11, 95% CI 0.23 to 5.40, one study, 339 infants), 35 to 36 + 6 weeks (no events in 191 infants) and greater than 36 weeks (no events in 42 infants). Birthweight was significantly decreased in infants entering a trial from 30 to 32 + 6 weeks (FWMD -190.64 grams, 95% CI -359.98 to -21.30 grams, one study, 319 infants), but not from less than 26 weeks (FWMD 63.14 grams, 95% CI -607.37 to 733.65 grams, one study, 49 infants), 26 to 29 + 6 weeks (FWMD 26.41 grams, 95% CI -215.55 to 268.37 grams, one study, 261 infants), 33 to 34 + 6 weeks (FWMD -38.72 grams, 95% CI -172.29 to 94.85 grams, onestudy, 353 infants), 35 to 36 + 6 weeks(FWMD -13.57 grams, 95% CI -175.45 to 148.31 grams, one study, 194 infants) and greater than 36 weeks (FWMD 73.89 grams, 95% CI -270.89 to 418.67 grams, one study, 42 infants). No statistically significant differences between groups treated with antenatal corticosteroids and controls were seen for combined fetal and neonatal deaths or fetal deaths alone in the different subgroups of gestational age at trial entry examined. Antenatal corticosteroids versus placebo or no treatment (by presence or absence in protocol of weekly repeat doses of corti￾costeroid) Data were available by the presence or absence in the protocol of weekly repeat doses of corticosteroid if the mother remained undelivered for several of the primary outcomes that relate to the motherand fetus/neonate.Therewas no differencein effect ofcor￾ticosteroid treatment on chorioamnionitis, puerperal sepsis, com￾bined fetal and neonatal death, fetal death, neonatal death, RDS or cerebroventricular haemorrhage between studies which used a single course of antenatal corticosteroid and studies that allowed weekly repeats if the women remained undelivered. D I S C U S S I O N The results of the 21 studies included in this updated review cat￾egorically support the conclusion of the previous review (Crow￾ley 1996), that treatment with antenatal corticosteroids reduces neonatal death, respiratory distress syndrome (RDS), and cere￾broventricular haemorrhage in preterm infants. Furthermore, treatment with antenatal corticosteroids is not associated with changes in the rates of maternal death, maternal infection, fetal death, neonatal chronic lung disease or birthweight. Treatment with antenatal corticosteroids is also associated with a reduction in theincidence of neonatal necrotising enterocolitisand systemic infections in the first 48 hours of life, as well as a reduction in the need for respiratory support or neonatal intensive care unit admission. However, one trial (Amorim 1999) recruiting women with severe preeclampsia, using a protocol that included repeat weekly courses ofantenatal betamethasoneif thewomen remained undelivered, suggested that the treated women were at increased risk of gestational diabetes. The women in this trial had a fasting glucose tolerance test more than 72 hours after the initiation of the study treatment if they were undelivered; 123 (56%) women under went the glucose tolerancetest. It may not beappropriateto generalise this to women without pre-eclampsia. It is also difficult to determine whether the fact that the protocol in this study used weekly repeat courses of antenatal corticosteroids was of relevance to the outcome. Concern has been expressed astowhetherantenatalcorticosteroids are beneficial in the current era of advanced neonatal practice, on the basis that previous conclusions concerning their benefits were based mainly on data from the 1970s. This update shows thatcombined fetaland neonatal death, neonatal death, RDS and cerebroventricular haemorrhageareall significantly reduced in the subgroup of trialsconducted in the 1990s.Thesetrialscontributed 26% of the overall data to thereview. This supports thecontinued use of antenatal corticosteroids. The gestational age range at which antenatal corticosteroids pro￾vide benefit has been subject to debate, with some reviews sug￾gesting no benefit at less than 28 weeks (Crowley 1996). Previ￾ously the effect of antenatal corticosteroids has been examined by subgroups based on gestational age at delivery. This review shows that antenatal corticosteroids reduce the incidence of cerebroven￾tricular haemorrhage even in those infants born before 28 weeks. However, this review also examined outcomes by subgroups based on the clinically more relevant measure; gestational age at first dose of treatment (gestational age at trial entry). RDS is reduced when corticosteroids are first given at 26 to 29.9 weeks, 30 to 32.9 weeksand 33 to 34.9 weeks. Furthermore, both cerebroventricular haemorrhage and neonatal death are reduced at 26 to 29.9 weeks. No difference is shown for primary outcomes at gestational ages of less than 26 weeks. While eight trials included in this review recruited pregnancies from less than 26 weeks’ gestation, and a further three did not specify the lower gestational age for entry, only one trial (n = 49 infants) contributed data to this review at this extreme gestation. Antenatal corticosteroid use reduces neonatal death even when infants are born less than 24 hours after the first dose has been given. Reduction in RDS is seen in infants born up to seven days after the first dose. This review has not shown any benefit in pri￾Antenatal corticosteroids for accelerating fetal lung maturation for women at risk of preterm birth (Review) 12 Copyright © 2006 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd 第 106 页