Beneficial roles of sex steroids on the developing infant have been proposed from animal data and observational studies in humans. During pregnancy, the placenta is the primary site of estrogen secretion, utilising precursors that arise from both maternal and fetal compartments (Siler-Kohr 1992). Levels of unconjugated estrone and estradiol produced by the placenta are at least 100 times the levels found in the non-pregnant adult female (Kenny 1973) and the fetus is exposed to these high levels of estrogen following transplacental passage (Hercz 1985) . Following birth, this exposure to high levels of estrogen ceases within 24 hours. At term this state is physiological, ie. the normal development of organ systems progresses in the face of an acute withdrawal of estrogen at term. Placental progesterone synthesis is independent of fetal precursors, and most of the secreted progesterone enters the maternal circulation; however, fetal plasma concentrations are also significantly elevated (Tulchinsky 1994). The preterm infant is deprived of this elevated sex steroid exposure through to term, and the biological effects of this deprivation on developing organ systems are unknown.
Observation of preterm infants receiving either breast milk or pre-term formula has revealed that even when supplemented with vitamin D, accumulation of calcium and phosphorous is inefficient by comparison to the fetus in-utero, possibly due to poor gut absorption (Lyon 1984). Given the relative deficiency of sex steroids in the preterm infant, it has been postulated that replacement of these hormones may facilitate mineral accretion (Trotter 2002a) . Human osteoblasts express estrogen receptors and estrogens have been shown to stimulate osteoblast proliferation. Estradiol may also prevent glucocorticoid-induced apoptosis of osteoblasts (Trotter 2000). In addition to a potential role early in life, estrogen clearly plays a protective role in post-menopausal osteoporosis (Lindsay 1988).
Animal studies have demonstrated that administration of estrogen to pregnant rabbits increases surfactant production by the fetus (Chu 1985), and accelerates morphologic maturation of the fetal lung (Khosla 1981). Whether or not sex steroid deficiency contributes to the pathogenesis of surfactant deficiency in the preterm infant is speculative.
It has also been postulated that the sex steroids have a role to play in the central nervous system. Two possible mechanisms exist for such a role. The first is in the minimisation of brain injury following an ischaemic event - a theory which derives plausibility from the observation of improved outcome in stroke patients who received estrogen replacement following their injury (Melton 2001). The second relates to a potential role of estrogens in the sequence of normal brain development (Beyer 1999). Specifically, estrogens have been shown to promote synapse formation (Matsumoto 1991) and progesterone has been shown to promote myelination (Baulieu 1996). Again, speculation exists about the role of these hormones on neurodevelopment in the human infant. There is evidence from observational studies of girls with Turner syndrome that estrogen therapy can improve verbal and nonverbal memory (Ross 2000) and that women with Turner syndrome have altered mental health (Downey 1989). Estrogen levels modulate the activity of platelet activating factor acetyl hydrolase, and platelet activating factor has been implicated in the pathogenesis of necrotising enterocolitis (Singh 2001). These links, whilst tenuous, raise biologically plausible possibilities about the potential role of sex steroids on various aspects of the preterm infant's wellbeing.
Administration of sex steroids is not without risk. In the 1960's, women with high risk pregnancies were treated with diethylstilbestrol (DES). Epidemiological studies have since demonstrated strong associations between such therapy and abnormalities in the offspring of these pregnancies. Specifically, in the daughters of treated women there is a significantly increased risk of vaginal adenocarcinoma (Herbst 1975) and of unfavourable outcome of pregnancies (Barnes 1980). Sons of treated women have also been shown to have abnormalities of the urogenital tract, specifically urethral stricture (Henderson 1976).
Maintenance of plasma levels of estradiol and progesterone similar to those measured in-utero have been achieved following the intravenous administration of these hormones (Trotter 1999). Postnatal treatment of the preterm infant with estradiol and progesterone resulted in the attainment of measures of uterine width similar to those of the fetus in-utero, compared to untreated infants whose uteri were significantly smaller than either fetuses of the same post menstrual age, or postnatally treated infants (Trotter 2002a). Whilst biological activity seems attainable, clinically important effects have yet to be established.
To determine if estrogens or progestins, either alone or in combination, when compared to placebo or no treatment, reduce morbidity and/or mortality in preterm infants. Subgroup analyses will be performed as follows where the trial data allow. (1) To determine the effects of estrogen alone or progestin alone, or combined estrogen and progestin therapy according to gestation, (2) the effect of the dose and duration of treatment, (3) the effect of the route of administration of treatment and (4) gender of study participants.
Secondary outcome measures will include any of the following:
1. Length of stay in
The standard search strategy of the Neonatal Review Group as outlined in the Cochrane Library (Issue 2, 2004) was used. This included searches of the Oxford Database of Perinatal Trials, Cochrane Controlled Trials Register (The Cochrane Library, Issue 2, 2004), MEDLINE, previous reviews including cross references, abstracts, conferences and symposia proceedings (Perinatal Society of Australia and New Zealand 1998-2004 and Pediatric Academic Societies meetings 1998-2004).
The search of MEDLINE, 1966 to July 2004 inclusive, included MeSH searches using the following terms ("infant, premature" AND [progest$ OR oestr$]) and text searches using the terms "[progest$ and oestr$]". Searches were limited to "clinical trials". No language restrictions were applied.
This systematic review followed the Cochrane Collaboration methodology according to guidelines of the Neonatal Review Group. Three reviewers independently identified the studies which are included, assessed the quality of the studies and extracted the data. Agreement between reviewers was reached through discussion. Methodological quality assessment was based on 1) blinding of randomisation, 2) blinding of intervention, 3) completeness of follow-up and 4) blinding of outcome measurement. When necessary, additional information and clarification of published data was requested from the authors of individual trials. Meta-analyses were performed using the fixed effects model, and a heterogeneity test (I-squared test) was applied to ensure that pooling of data was valid. Relative risk (RR) and risk difference (RD) were calculated for dichotomous data and weighted mean difference (WMD) for continuous data, with 95% confidence intervals (CI) for all analyses. The number needed to treat (NNT) and associated 95% CI were determined for a statistically significant reduction in the RD.
When extracting data from the included studies for the secondary outcome chronic lung disease, variations in definition were accepted as a post facto departure from the original protocol.
Two relevant studies were identified by the search.
The study by Trotter 1999a, included in this review, reported a cohort of infants that were reported in several different publications. Preliminary pharmacokinetic data relating to this study were reported in Trotter 1999. No clinically relevant outcomes were reported in that paper. Various aspects of clinical outcome for this cohort were then reported in Trotter 2002a and Trotter 2002b. Neurodevelopmental follow-up at 15 months of age was reported in Trotter 2001. These reports will be collectively referred to hereafter as Trotter 1999a. Female infants with gestational age < 29 weeks and birthweight < 1000grams were eligible for this study. During the study period, there were 40 eligible infants and consent was obtained for 30 infants, with 15 infants enrolled into either a replacement or control group. Infants in the replacement group received 15 ml/kg/day continuous intravenous infusion of estradiol and progesterone lipid 5% mixture, while the control group received an estradiol-progesterone free lipid 5% mixture. The dosage of estradiol and progesterone was tailored for each individual infant based on measurements of hormones in cord blood and subsequent blood samples at day 1, 3, 5, 7 and then weekly until the seventh postnatal week. The investigators' aim was to maintain plasma levels of 2000-6000 pg/ml for estradiol and 300-600 ng/ml for progesterone, corresponding to intrauterine levels at the same gestation. Once intravenous access was no longer required, the replacement group received estradiol and progesterone transepidermally for a total intervention period of six weeks. Refer to Table of Characteristics of Included Studies.
There was one other randomised trial of estrogen replacement reported by Shanklin 1970. This study was conducted between 1964 and 1966. In this study, 'premature' infants, not defined by gestation or birthweight, received an intramuscular injection within one hour of birth of either equine estrogens or lactose in diluent. The vials had been randomised by one of the investigators, their labels removed and then numbered. In a table of results, only 16 of 287 infants included in this study had birthweights less than 1000 grams. On the other hand, 187 of 287 (65%) of included infants had a birthweight over 2001 grams, effectively excluding them from the gestational range of interest in this review. This study was therefore excluded (see Table of Characteristics of Excluded Studies).
Randomisation: Eligible infants in the study of Trotter 1999a were randomly assigned in blocks of four to receive either estradiol and progesterone replacement or placebo (controls). Method of randomisation was unstated in all reports of this study. Sealed envelopes were used.
Allocation concealment: Allocation to replacement or control group was carried out with the use of sealed envelopes.
Blinding of intervention: Whilst intravenous access was required,
treatment infants received 15ml/kg/day of estradiol and progesterone-lipid
5% mixture whereas control infants received 15ml/kg/day of an estradiol-
and progesterone-free lipid 5% mixture. Dosage of estradiol and progesterone
was individualised for subjects in the replacement group, and for some infants
this involved a change in infusion rate. There was no change in rate of infusion
reported for the placebo group, hence potentially unblinding the intervention.
Once intravenous access was no longer required, replacement group infants
received estradiol and progesterone ointment transepidermally. There was
no placebo equivalent reported for transepidermal administration of the intervention.
Intervention was conducted for six weeks in total. The number of infants
in either group requiring intravenous therapy for the first six weeks postnatal
life is not reported.
Blinding of outcome measurement: In the reports of Trotter 1999a and Trotter 2002a no reference is made to blinding of outcome measurement. In the report of Trotter 2002b the investigator assessing vaginal cytology was blind to intervention. There was no blinding of any other short term outcome measurement. Importantly, in the report of Trotter 2001, the psychologist and paediatric neurologist performing assessments of long term neurodevelopmental outcome were blind to treatment group.
Completeness of follow-up: The report of Trotter 2001 states that 25 survivors of 30 infants randomised to replacement or placebo were eligible for long-term follow-up. 24 infants were subsequently examined, and Bayley results were reported for 21 infants (21/25 =84%). Three infants (two receiving sex steroid and one control) were excluded from this analysis because sensory impairment invalidated administration of the Bayley scales, or the infant was too tired to perform the tests.
Thirty female infants in total were randomised in the study of Trotter 1999a to receive either combined estradiol and progesterone replacement or placebo for the first six postnatal weeks.
PRIMARY OUTCOME MEASURES:
Neonatal mortality and mortality to hospital discharge: There
was no significant difference in either neonatal mortality (RR 2.0, 95%CI
0.2 to 19.78) or mortality to hospital discharge (RR 0.67, 95%CI 0.13 to
3.44) between preterm female infants who received estradiol and progesterone
and those who received placebo.
Neurodevelopmental impairment: Trotter 2001report neurodevelopmental outcome at 15 months of age for 24 of the 30 subjects first reported in Trotter 1999a. They report a trend to reduction of psychomotor index on Bayleys Developmental assessment for control infants compared to those that received hormone replacement, but only median, minimum and maximum scores are reported. Those infants receiving estradiol and progesterone had median MDI scores (min, max) of 89 (71, 107) compared to controls who had median MDI 93 (49, 111). Steroid supplemented infants had median PDI (min, max) of 101 (49, 121) compared to controls who had median PDI 71 (49, 121). The report of Trotter 2001 also suggests that control infants had more neurological impairment than those that received hormone replacement. From the data presented, there were no significant differences in neurodevelopmental outcome between the two groups. There are no data allowing comparison of components of neurological impairment (blindness, hearing impairment and cerebral palsy).
SECONDARY OUTCOME MEASURES:
1. Length of stay in newborn intensive care is not reported. Trotter 1999a
report a reduction in median length of stay in hospital for those infants
who received estradiol and progesterone replacement compared to controls,
however further analysis is not possible from the data available.
2. Evidence of organ dysfunction
Planned subgroup analyses could not be performed. There was only one trial enrolling only female subjects and randomising them to receive both estradiol and progesterone or placebo. Thus effects of gender or of individual hormone replacement could not be addressed. Administration of hormone was initially intravenous and then transdermal once intravascular access was no longer clinically required, to a total of six weeks of hormone treatment. Analysis of differences based on route of administration was not possible.
This systematic review reports the results of 30 preterm female infants, with gestational age less than 29 weeks and birthweight less than 1000 grams, who were randomised to receive either a combination of estradiol and pprogesterone intravenously and transepidermally, or placebo intravenously, for the first six weeks of postnatal life. The dosage of hormone administered was individually tailored to maintain plasma estradiol and progesterone levels similar to those found in-utero at the same post-conceptional age. Despite the potential benefits of sex steroid administration, no clinically significant benefits are detectable from the administration of a combination of estradiol and progesterone, although the number of infants studied to date is small. The study reported by Trotter 1999a is described as a pilot study justifying the small number of infants randomised. To date there has been no data to suggest that some infants will maintain hormone levels similar to the in-utero status without supplementation, nor have there been any studies examining the effect of individual hormone replacement, different routes of administration or hormone replacement in male preterm infants. The dramatic difference in hormone profiles between the fetus in-utero compared to ex-utero infants of the same post-conceptional age, coupled with the limited findings suggesting benefits of sex steroid replacement, leave this area open to further scientific evaluation.
| Study | Methods | Participants | Interventions | Outcomes | Notes | Allocation concealment |
| Trotter 1999a | Random study: yes, method not stated. Allocation concealment: sealed envelopes batched into groups of four. Blinding of intervention: Initially, continuous intravenous infusions were commenced at a rate of 15ml/kg/day of either estradiol-progesterone-lipid 5% mixture for the replacement group or lipid 5% mixture alone for the control group. Dosage of estradiol and progesterone varied between subjects in the replacement group, and this was varied for some infants by changing the infusion rate or altering the concentration of hormone in the infusate. No changes in infusion rate were reported for the placebo group - thus potentially unblinding the intervention. Once intravenous access was no longer required, estradiol and progesterone were administered transepidermally with ointment. There is no mention in any of the reports that a transepidermal placebo was applied to control infants. Blinding of measurement: variably reported. Neurodevelopmental outcome data were collected by a psychologist and paediatric neurologist blind to treatment group. Losses to follow-up: yes, Neurodevelopmental follow-up data available for 24/25 survivors. Allocation concealment: adequate. | Female infants with gestational age < 29 weeks and birthweight < 1000g were eligible. During the study period, consent was obtained for 30 of 40 eligible infants, and 15 infants were randomised to each of replacement and control groups. | Intervention commenced as soon as intravenous access was established on the first day of life. In the replacement group, a continuous infusion of estradiol-progesterone-lipid 5% mixture was commenced at 15 ml/kg/day, and the infusion rate or concentration of hormone in the infusate adjusted to achieve plasma hormone levels of 2000-6000 pg/ml of estradiol and 300-600 ng/ml of progesterone. Dosage adjustment was based on measurement of plasma hormones from cord blood, and then serial measurements on days 1, 3, 5, 7 and then weekly until the seventh postnatal week. Once intravenous access was no longer required, estradiol and progesterone were administered transepidermally with ointment. No transepidermal placebo is reported. | Short term morbidity (intraventricular haemorrhage, retinopathy of prematurity, necrotising enterocolitis and chronic lung disease) and mortality. Calcium and phosphorus balance are reported for a subset of the cohort (11 replacement infants and 9 controls). Uterine and mammary growth, and vaginal cytology are reported for the 25 survivors (13 replacement and 12 controls) of the original 30 randomised subjects. Neurodevelopmental follow-up is reported for 24 of 25 survivors. | A |
| Study | Reason for exclusion |
| Shanklin 1970 | 'Prematurity' not defined for this study and only 16/287 included infants had a birthweight < 1000 grams, whereas 187/287 infants had a birthweight > 2000 grams. Thus most infants studied would have had gestations > 30 weeks. |
Trotter A, Bokelmann B, Sorgo W, Bechinger-Kornhuber D, Heinemann H, Schmucker G et al. Follow-up examination at the age of 15 months of extremely preterm infants after postnatal estradiol and progesterone replacement. Journal of Clinical Endocrinology and Metabolism 2001;86:601-3.
Trotter A, Maier L, Grill H-J, Wudy SA, Pohlandt F. 17-Beta-estradiol and progesterone supplementation in extremely low-birth-weight infants. Pediatric Research 1999;45:489-93.
* Trotter A, Maier L, Grill HJ, Kohn T, Heckmann M, Pohlandt F. Effects of postnatal estradiol and progesterone replacement in extremely preterm infants. Journal of Clinical Endocrinology and Metabolism 1999;84:4531-5.
Trotter A, Maier L, Kohn T, Bohm W, Pohlandt F. Growth of the uterus and mammary glands and vaginal cytologic features in extremely premature infants with postnatal replacement of estradiol and progesterone. American Journal of Obstetrics and Gynecology 2002;186:184-8.
Trotter A, Maier L, Pohlandt. Calcium and phosphorus balance of extremely preterm infants with estradiol and progesterone replacement. American Journal of Perinatology 2002;19:23-9.
Shanklin DR, Wolfson SL. Aqueous estrogens in the management of respiratory distress syndrome. Journal of Reproductive Medicine 1970;5:53-71.
* indicates the primary reference for the study
Barnes AB, Colton T, Gundersen J, Noller KL, Tilley BC, Strama T, Townsend DE, Hatab P, O'Brien PC. Fertility and outcome of pregnancy in women exposed in utero to diethylstilbestrol. New England Journal of Medicine 1980;302:609-13.
Baulieu EE, Schumacher M, Koenig H, Jung-Testas I, Akwa Y. Progesterone as a neurosteroid: actions within the nervous system. Cellular and Molecular Neurobiology 1996;16:143-54.
Beyer C. Estrogen and the developing mammalian brain. Anatomy and Embryology 1999;199:379-90.
Chu AJ, Rooney SA. Estrogen stimulation of surfactant synthesis. Pediatric Pulmonology 1985;1[suppl]:S110-14.
Downey J, Ehrhardt AA, Gruen R, Bell JJ, Morishima A. Psychopathology and social functioning in women with Turner syndrome. Journal of Nervous and Mental Disease 1989;177:191-201.
Henderson BE, Benton B, Cosgrove M, Baptista J, Aldrich J, Townsend D, Hart W, Mack TM. Urogenital tract abnormalities in sons of women treated with diethylstilbestrol. Pediatrics 1976;58:505-7.
Herbst AL, Poskanzer DC, Robboy SJ, Friedlander L, Scully RE. Prenatal exposure to stilbestrol. A prospective comparison of exposed female offspring with unexposed controls. New England Journal of Medicine 1975;292:334-9.
Hercz P. Quantitative changes in steroid and peptide hormones in the maternal-fetoplacental system between the 28th - 40th weeks of pregnancy. Acta Medica Hungarica 1985;42:29-39.
Kenny FM, Angsusingha K, Stinson D, Hotchkiss J. Unconjugated estrogens in the perinatal period. Pediatric Research 1973;7:826-31.
Khosla SS, Walker Smith GJ, Parks PA, Rooney SA. Effects of estrogen on fetal rabbit lung maturation: Morphological and biochemical studies. Pediatric Research 1981;15:1274-81.
Lindsay R. Sex steroids in the pathogenesis and prevention of osteoporosis. In: Melton LJ, Riggs L, editor(s). Osteoporosis: Etiology, Diagnosis and Management. New York: Raven Press, 1988:333-358.
Lyon AJ, McIntosh N. Calcium and phosphorus balance in extremely low birthweight infants in the first six weeks of life. Archives of Disease in Childhood 1984;59:1145-50.
Matsumoto A. Synaptogenic action of sex steroids in developing and adult neuroendocrine brain. Psychoneuroendocrinology 1991;16:25-40.
Melton L. What can sex hormones do for the damaged brain? Lancet 2001;358:818.
Ross JL, Roeltgen D, Feuillan P, Kushner H, Cutler GB. Use of estrogen in young girls with Turner syndrome: effects on memory. Neurology 2000;54:164-70.
Siler-Khodr TM. Endocrine and paracrine function of the human placenta. In: Polin RA, Fox WW, editor(s). Fetal and neonatal physiology. Philadelphia, PA: WB Saunders, 1992:74-85.
Singh K, Caplan M, Moya FR. Effects of antenatal and postnatal steroids on platelet activating factor acetylhydrolase activity in preterm infants. Pediatric Research 2001;49:343A.
Trotter A, Maier L, Grill H-J, Wudy SA, Pohlandt F. 17-Beta-estradiol and progesterone supplementation in extremely low-birth-weight infants. Pediatric Research 1999;45:489-93.
Trotter A, Pohlandt F. The replacement of oestradiol and progesterone in very premature infants. Annals of Medicine 2000;32:608-14.
Trotter A, Bokelmann B, Sorgo W, Bechinger-Kornhuber D, Heinemann H, Schmucker G et al. Follow-up examination at the age of 15 months of extremely preterm infants after postnatal estradiol and progesterone replacement. Journal of Clinical Endocrinology and Metabolism. 2001;86:601-3.
Trotter A, Maier L, Pohlandt F. Calcium and phosphorus balance of extremely preterm infants with estradiol and progesterone replacement. American Journal of Perinatology 2002;19:23-9.
Trotter A, Maier L, Kohn T, Bohm W, Pohlandt F. Growth of the uterus and mammary glands and vaginal cytologic features in extremely premature infants with postnatal replacement of estradiol and progesterone. American Journal of Obstetrics and Gynecology 2002;186:184-8.
Tulchinsky D, Little AB. Maternal-Fetal Endocrinology. Second edition edition. Philadelphia, Pensylvania: W B Saunders Company, 1994.
01.01 Neonatal mortality (to 28 days of age)
01.02 Mortality to hospital discharge
01.03 Endotracheal intubation
01.04 Chronic lung disease amongst survivors
01.05 Necrotising enterocolitis
| Comparison or outcome | Studies | Participants | Statistical method | Effect size |
|---|---|---|---|---|
| 01 Estradiol and progesterone replacement versus placebo | ||||
| 01 Neonatal mortality (to 28 days of age) | 1 | 30 | RR (fixed), 95% CI | 2.00 [0.20, 19.78] |
| 02 Mortality to hospital discharge | 1 | 30 | RR (fixed), 95% CI | 0.67 [0.13, 3.44] |
| 03 Endotracheal intubation | 1 | 30 | RR (fixed), 95% CI | 0.92 [0.62, 1.36] |
| 04 Chronic lung disease amongst survivors | 1 | 27 | RR (fixed), 95% CI | 0.12 [0.01, 2.02] |
| 05 Necrotising enterocolitis | 1 | 30 | RR (fixed), 95% CI | 0.33 [0.01, 7.58] |
A/Prof Terrie E Inder
Peadiatric Neurologist/Neonatologist
Neonatal Services Division
Royal Women's Hospital
132 Grattan Street
Carlton, Melbourne
Victoria AUSTRALIA
3053
Telephone 1: +61 3 93442000 extension: 3178
Facsimile: +61 3 93472731
E-mail: terrie.inder@rch.org.au
| This review is published as a Cochrane review in The
Cochrane Library, Issue 4, 2004 (see http://www.thecochranelibrary.com/ for information).
Cochrane reviews are regularly updated as new evidence emerges and in response
to comments and criticisms, and The Cochrane Library should be consulted
for the most recent version of the Review. |
