Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 7-Day Trial for You or Your Team.

Learn More →

STRIDER: Sildenafil therapy in dismal prognosis early-onset intrauterine growth restriction – a protocol for a systematic review with individual participant data and aggregate data meta-analysis and trial sequential analysis

STRIDER: Sildenafil therapy in dismal prognosis early-onset intrauterine growth restriction – a... Background: In pregnancies complicated by early-onset extreme fetal growth restriction, there is a high risk of preterm birth and an overall dismal fetal prognosis. Sildenafil has been suggested to improve this prognosis. The first aim of this review is to assess whether sildenafil benefits or harms these babies. The second aim is to analyse if these effects are modified in a clinically meaningful way by factors related to the women or the trial protocol. Methods/Design: The STRIDER (Sildenafil Therapy In Dismal prognosis Early-onset intrauterine growth Restriction) Individual Participant Data (IPD) Study Group will conduct a prospective IPD and aggregate data systematic review with meta-analysis and trial sequential analysis. The STRIDER IPD Study Group started trial planning and funding applications in 2012. Three trials will be launched in 2014, recruiting for three years. Further trials are planned to commence in 2015. The primary outcome for babies is being alive at term gestation without evidence of serious adverse neonatal outcome. The latter is defined as severe central nervous system injury (severe intraventricular haemorrhage (grade 3 and 4) or cystic periventricular leukomalacia, demonstrated by ultrasound and/or magnetic resonance imaging) or other severe morbidity (bronchopulmonary dysplasia, retinopathy of prematurity requiring treatment, or necrotising enterocolitis requiring surgery). The secondary outcomes are improved fetal growth velocity assessed by ultrasound abdominal circumference measurements, gestational age and birth weight (centile) at delivery, and age-adequate performance on the two-year Bayley scales of infant and toddler development-III (composite cognitive score and composite motor score). Subgroup and sensitivity analyses in the IPD meta-analysis include assessment of the influence of several patient characteristics: an abnormal or normal serum level of placental growth factor, absent/reversed umbilical arterial end diastolic flow at commencement of treatment, and other patient characteristics available at baseline such as gestational age and estimated fetal weight. The secondary outcomes for mothers include co-incidence and severity of the maternal syndrome of pre-eclampsia, mortality, and other serious adverse events. Discussion: Trials are expected to start in 2013–2014 and end in 2016–2017. Data analyses of individual trials are expected to finish in 2019. Given the pre-planned and agreed IPD protocol, these results should be available in 2020. Keywords: Adverse neonatal outcome, Fetal growth restriction, Individual participant data meta-analysis, Sildenafil * Correspondence: [email protected] Department of Obstetrics and Gynecology, Academic Medical Centre, Room: H4-278, PO BOX 22660, 1100 DD Amsterdam, The Netherlands Full list of author information is available at the end of the article © 2014 Ganzevoort et al.; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Ganzevoort et al. Systematic Reviews 2014, 3:23 Page 2 of 9 http://www.systematicreviewsjournal.com/content/3/1/23 Background the belief that rest will enhance the uteroplacental circula- Severe early-onset fetal growth restriction (FGR) compli- tion at the expense of that to the glutei and quadriceps cates approximately 0.4% of pregnancies and severely in- muscles. creases the risk of perinatal morbidity and mortality. This There is evidence from ex vivo and animal models of is particularly due to premature delivery, both for fetal growth restriction that the phosphodiesterase 5 inhibitor and for secondary maternal indications such as the devel- sildenafil citrate increases average pup birth weight and opment of pre-eclampsia [1]. Placental disease, conse- improves uteroplacental blood flow (umbilical artery, quent on deficient uteroplacental blood flow, includes uterine artery) [10-12]. Sildenafil citrate may thus offer a FGR, pre-eclampsia, and placental abruption and has been potential therapeutic strategy to improve uteroplacental implicated in more than 50% of iatrogenic premature blood flow in human pregnancies complicated by severe births [2]. For this reason, the problem of severe FGR FGR [12]. There is evidence, albeit limited, to support forms a substantial portion of the population that tertiary the use of sildenafil in human pregnancy. A small trial in care centres care for. women with early-onset pre-eclampsia (a maternal pres- The effect of severe early-onset FGR is particularly sig- entation of placental compromise) showed no resolution nificant: of those born alive, less than a third will survive of pre-eclampsia, but an effect on birth weight [13]. In their neonatal intensive care unit (NICU) stay without addition, a small cohort study showed a tendency to- significant neurodevelopmental sequelae. Survival rates wards more live-born children that survived intact to pri- for severely growth-restricted fetuses very remote from mary discharge after sildenafil intervention [14]. From all term (<28 weeks’ gestation) vary from 7% to 33% [1,3,4]. observations thus far, there are no indications that sildena- As these early-onset FGR children are born very pre- fil in the second trimester has significant fetotoxicity. term, there are significant risks of neonatal mortality, Coordinated initiatives have been launched and ap- major and minor morbidity, and long-term health seque- proved/shortlisted for funding in networks in the United lae [5,6]. These are not only strongly gestational-age re- Kingdom/Ireland, The Netherlands, and New Zealand/ lated [7], but are also related to FGR [8]. Although most Australia. Initiatives in Canada and the United States of of the surviving children are without severe disabilities America are in an earlier phase. These trials are embed- at school age, many of them meet serious difficulties in ded in a global network (Global Obstetrics Network – everyday life and the burden of mild developmental ab- GONet) [15]. This network was initiated to provide a normalities and behavioural and learning disorders in- forum for international interaction and foster communi- creases with age. In adolescents, as many as 40% of the cation between groups that perform clinical studies in survivors will not be able to become fully independent obstetrics. GONet has started the collaborative initiative adults [9]. to resolve the underlying question: does sildenafil versus The direct costs of care include the increased cost of placebo improve outcomes in early-onset severely growth intensive antenatal surveillance (with or without a period restricted fetuses with a dismal prognosis? Within each of hospitalisation), caesarean delivery (if considered vi- part of the network, stand-alone randomised clinical trials able), NICU care, routine post-NICU follow-up, need for will be conducted; a prospective systematic review with special schooling and educational support (speech ther- meta-analyses and trial sequential analyses of individual apy, physical therapy, psychological support), and, often, participant data (IPD) and aggregate data will then be per- specialised neurodevelopmental assessments and inter- formed. The concerted effort in the design and conduct of ventions. Such costs are in addition to the indirect costs these large international studies provides an efficient use to families (grief, lost wages, supportive services, and re- of scarce resources and will optimise resolution of the re- lationship stress) and represent a considerable economic search questions, in addition to facilitating adequate burden. Safe pregnancy prolongation accompanied by power for subgroup analyses. additional fetal growth would be associated with signifi- cant societal and financial savings. Methods/Design The use of ultrasound Doppler waveform analysis in Study design pregnancies complicated by FGR suggests compromised Prospectively planned systematic review with IPD and uteroplacental circulation and placental hypoperfusion. aggregate data meta-analysis [16] and trial sequential Currently there are no specific evidence-based therapies analysis [17-21] of randomised clinical trials comparing for placental insufficiency and early-onset severe FGR. sildenafil with matching placebo tablets or no interven- Non-specific interventions include primarily lifestyle mod- tion for treatment of severe early-onset FGR. ifications, such as reducing or stopping work, stopping aerobic exercise, rest at home, and hospital admission for Inclusion and exclusion criteria for the studies rest and surveillance. These interventions, which are not Studies, published or unpublished, will be included if supported by evidence from randomised trials, are used in they are randomised clinical trials. Other study designs Ganzevoort et al. Systematic Reviews 2014, 3:23 Page 3 of 9 http://www.systematicreviewsjournal.com/content/3/1/23 for assessment of benefits will not be included. For the join the STRIDER Study Group have agreed a common assessment of harm we will also include observational core data set and have agreed to supply their trial’s IPD studies. Given the pre-planned nature of the IPD meta- to a central repository. Randomisation will be centrally analysis, most or all studies will have been identified be- controlled using an on-line computerised randomisation fore the start of the individual trials and will participate service. All have agreed to collect all relevant variables in central data collection. If other trials, yet unknown, on baseline characteristics and outcomes (the primary are conducted simultaneously, their eligibility will be and secondary outcomes of this protocol). All have assessed independently and unblinded for author and agreed to centrally collect data. The University of British journal by two members of the Sildenafil Therapy In Columbia (UBC) will host data management through Dismal prognosis Early-onset intrauterine growth Re- the UBC Perinatal Clinical Trials Unit. Wherever pos- striction (STRIDER) IPD Study Group. Any differences sible, data will be collected as continuous rather than in opinion regarding eligibility will be resolved by dis- dichotomous measurements. All trialists have agreed cussion. If the study design is suboptimal for IPD meta- to collect information on the number of eligible but analysis or the authors do not cooperate, the study will not-included patients. If the patient consents, a number be included for aggregate data analysis. of relevant baseline and outcome parameters will be We will use the Cochrane Central Register of Con- collected. trolled Trials (CENTRAL), MEDLINE, Embase, LILACS, If other relevant randomised clinical trials are identi- and Science Citation Index Expanded using the terms fied, data collected from those studies will include all “fetal growth retardation” and “sildenafil”. In addition, women randomised and coded for anonymity (date of we will access the WHO Trial Portal [22], Current Con- birth, centre identification), baseline data for descriptive trolled Trials [23], the Australian and New Zealand Trials purposes and analyses, details of the intervention given Register [24], and the ISRCTN register [25] to identify re- (date of randomisation, allocated intervention), and out- cently completed or ongoing trials. Experts in the field comes, to allow planned analyses. Trialists will provide and trialists will be asked if they know of any unpublished non-identified IPD in any convenient format by encrypted, or other trials. The currently known trials that will be con- electronic transfer where possible or other means as ducted and included are: needed. The individual trial data will be recoded as re- quired, and stored in the secure database that will only be UK/Ireland: Planned sample size 112; primary accessible by authorised personnel of the UBC Perinatal outcome: prolongation of gestational age at birth. Clinical Trials Unit. Trialists will be asked to verify all The Netherlands: Planned sample size 354; primary recoded data prior to any analysis and the data will not be outcome: infant survival to term age without evidence used for any other purpose without permission of all of serious adverse neonatal outcome. collaborators. New Zealand/Australia: Planned sample size 122; The data will be checked with respect to range, in- primary outcome: fetal growth velocity. ternal consistency, missing or extreme values, errors, Canada: Planned sample size 90; primary outcome: and consistency with published reports. Trial details, fetal growth velocity. such as randomisation methods, blinding, and interven- tion details, will be crosschecked against published reports, Types of participants trial protocols, and data collection sheets. Inconsistencies Women with a singleton pregnancy between 20 + 0 and or missing data will be discussed with the individual trial- 29 + 6 weeks with severe fetal growth restriction of likely ists and attempts will be made to resolve any problems by placental origin, and with estimated significant likeli- consensus. hood of fetal/neonatal death. Data items to be collected Type of intervention Trial level information Sildenafil versus placebo tablet orally or versus no inter- vention until fetal death, delivery, or 32 weeks of gesta- 1. Dates the trial opened and closed accrual. tion (whichever comes first). Co-interventions will be 2. Number of participants randomised. allowed provided they are administered similarly in the 3. Informed consent procedures. intervention groups. 4. Methods of random allocation. 5. Stratification factors (if any). Data collection and management 6. Methods of allocation concealment. The investigators of each individual trial have designed 7. Blinding procedures for outcome assessment. each trial together before launch and patient recruit- 8. Details of the planned intervention in the drug arm: ment. All of the listed eligible trials that have agreed to frequency, timing, doses. Ganzevoort et al. Systematic Reviews 2014, 3:23 Page 4 of 9 http://www.systematicreviewsjournal.com/content/3/1/23 9. Details of the planned intervention in the placebo 2. Date and time of delivery. arm: frequency, timing, doses. 3. Birth weight, 5 minute Apgar scores, sex. 4. Mode of delivery. Participant-level information – maternal characteristics at 5. Neonatal morbidity. study entry 6. Date and time if neonatal death. 7. Childhood follow-up assessment. 1. Unique identification coded for anonymity. 8. Adverse events and adverse effects for the neonate. 2. Maternal age. 3. Body mass index. Participant-level information – data on actual trial intervention 4. Parity. 5. Abdominal circumference. 1. Type of treatment given (placebo, sildenafil, dose). 6. Ethnicity. 2. Number of doses actually given. 7. Hypertensive disease (type of). 3. Interval between first dose and 32 weeks or delivery 8. Uterine artery notching/pulsatility index. or fetal death. 9. Placental growth factor levels in maternal serum. 4. Total actual drug exposure. 10.Treatment (e.g., aspirin, low molecular weight heparin, oral steroids, insulin). Bias risk assessments This project will use the Cochrane Collaboration methods Participant-level information – fetal characteristics at study entry for assessing risk of bias within individual trials [26]: 1. Estimated due date. Allocation sequence generation 2. Date and time of randomisation. 3. Estimated fetal weight, abdominal circumference. Low risk of bias: sequence generation was achieved 4. Absent or reversed end-diastolic flow of the umbil- using computer random number generation or a ical artery, pulsatility index. random number table. Drawing lots, tossing a coin, 5. Pulsatility index of middle cerebral artery. shuffling cards, and throwing dice are adequate if performed by an independent person not otherwise Participant-level information – maternal information after involved in the trial. trial entry Uncertain risk of bias: the method of sequence generation was not specified. 1. Date and time of prenatal corticosteroid treatment. High risk of bias: the sequence generation method 2. Hypertensive disease (type of). was not random. 3. Highest blood pressure measured. 4. Proteinuria (maximum level recorded). Allocation concealment 5. Platelet count (lowest recorded). 6. Adverse events and adverse effects for the woman at Low risk of bias: the participant allocations could not time of treatment and post-partum. have been foreseen in advance of, or during, enrolment. Allocation was controlled by a central and independent Participant-level information – fetal information after trial entry randomisation unit. The allocation sequence was unknown to the investigators (for example, if the 1. Sequential assessments of the umbilical artery: absent allocation sequence was hidden in sequentially or reversed end-diastolic flow, pulsatility index. numbered, opaque, and sealed envelopes). 2. Sequential assessments of estimated fetal weight, Uncertain risk of bias: the method used to conceal the abdominal circumference. allocation was not described so that intervention 3. Sequential assessments of the middle cerebral artery: allocations may have been foreseen in advance of, or pulsatility index. during, enrolment. 4. Estimated date and time if fetal death. High risk of bias: the allocation sequence was likely to be 5. Adverse events and adverse effects for the fetus. known to the investigators who assigned the participants. Participant-level information – neonatal information after Blinding of participants, personnel, and outcome assessors trial entry Low risk of bias: blinding was performed adequately or 1. Unique baby information and mother identification the assessment of outcomes was not likely to be coded for anonymity. influenced by lack of blinding. Ganzevoort et al. Systematic Reviews 2014, 3:23 Page 5 of 9 http://www.systematicreviewsjournal.com/content/3/1/23 Uncertain risk of bias: there was insufficient High risk of bias: the trial is sponsored by the information to assess whether blinding was likely to industry. induce bias on the results. High risk of bias: no blinding or incomplete blinding, Other bias and the assessment of outcomes were likely to be influenced by lack of blinding. Low risk of bias: the trial appears to be free of other components that could put it at risk of bias. Incomplete outcome data Uncertain risk of bias: the trial may or may not be free of other components that could put it at risk of Low risk of bias: missing data were unlikely to make bias. treatment effects depart from plausible values. High risk of bias: there are other factors in the trial Sufficient methods, such as multiple imputation, have that could put it at risk of bias. been employed to handle missing data. Trials assessed as having ‘low risk of bias’ in all of the Uncertain risk of bias: there was insufficient individual domains specified in the review were information to assess whether missing data in considered ‘trials with low risk of bias’. Trials assessed combination with the method used to handle missing as having ‘uncertain risk of bias’ or ‘high risk of bias’ data were likely to induce bias on the results. in one or more of the specified in the review High risk of bias: the results were likely to be biased individual domains were considered trials with ‘high due to missing data. risk of bias’ [26-30]. Selective outcome reporting Dealing with missing data We intend to perform an intention-to-treat analysis when- Low risk of bias: all outcomes were pre-defined and re- ever possible. We will impute data for binary outcomes ported or all clinically relevant and reasonably expected using various scenarios, namely best-best analysis, worst- outcomes were reported. worst analysis, best-worst analysis, and worst-best analysis. Uncertain risk of bias: it was unclear whether all pre- For continuous outcomes, we will use available-participant defined and clinically relevant and reasonably ex- analysis. We will impute the standard deviation from P pected outcomes were reported. values according to the instructions given in the Cochrane High risk of bias: one or more clinically relevant and Handbook for Systematic Reviews of Intervention [26] and reasonably expected outcomes were not reported, and use the median for the meta-analysis when the mean is not data on these outcomes were likely to have been available. If it is not possible to calculate the standard devi- recorded. ation from the P value or the confidence intervals (CIs), we For a trial to be assessed with low risk of bias in the will impute the standard deviation as the highest standard selective outcome reporting domain, the trial should deviation in the other trials included under that outcome, have been registered either on the www.clinicaltrials. fully recognising that this form of imputation will decrease gov website or a similar register, or there should be a the weight of the trial for calculation of mean differences protocol, e.g., published in a paper journal. In the case and bias the effect estimate to no effect in the case of stan- when the trial was run and published in the years dardised mean difference. when trial registration was not required, we carefully scrutinised all publications reporting on the trial to Outcomes identify the trial objectives and outcomes. If usable The primary outcome is defined as: data on all outcomes specified in the trial objectives Neonatal survival until term age without serious neo- were provided in the publications results section, then natal morbidity defined as secondary outcomes 3–6. The the trial was considered low risk of bias trial in the outcome needs to be assessed formally at term age by Selective outcome reporting domain. clinic visit (or as a last-resort alternative by telephone). Secondary outcomes are: Industry bias 1. Any perinatal deaths (any know death irrespective of Low risk of bias: the trial appears to be free of industry cause). sponsorship that may manipulate the trial design, 2. Perinatal deaths in normally formed infants (absence conductance, or results of the trial. of severe congenital malformation, infection, or Uncertain risk of bias: the trial may or may not be genetic abnormality). free of for-profit bias as no information on clinical 3. Severe central nervous system injury (diagnosed by trial support or sponsorship is provided. ultrasound and/or magnetic resonance imaging) Ganzevoort et al. Systematic Reviews 2014, 3:23 Page 6 of 9 http://www.systematicreviewsjournal.com/content/3/1/23 periventricular leukomalacia grade II or more, or the data analyses. Any analyses conducted will be based intracerebral haemorrhage grade III or more, or on the checked and updated individual participant data hydrocephalus. from all available trials. All randomised participants with 4. Bronchopulmonary dysplasia (the requirement of outcome data available will be included in the analyses, ambulatory oxygen therapy >36 weeks corrected which will be performed on an intention-to-treat basis, gestational age). according to the treatment allocation at randomisation. 5. Retinopathy of prematurity (requiring treatment For each of the outcomes (as well as the individual such as laser therapy; grade 2/3 or more). components of each composite outcome), a one-stage 6. Necrotising enterocolitis (requiring surgery). approach to analysis will be taken so that the IPD from 7. Need for inotropes or vasopressors: number of days all eligible trials are included in a single model. Fitting a on inotropes or vasopressors, days on one, two, or single model for each outcome variable will enable the three different inotropes or vasopressors. variation across trials to be accounted for within the 8. Patent ductus arteriosus needing medical or surgical model by including a fixed trial effect. A treatment by treatment. trial interaction term will be tested to assess heterogeneity 9. Number of septic episodes, defined as culture of treatment effect across trials. If excessive statistical het- proven or clinical, with need for antibiotics for 5 or erogeneity in treatment effect or inconsistency across tri- more days. als is detected (that is, if the trial by treatment interaction 10. Total number of days on artificial ventilation. term is significant), then the rationale for combining 11. Postmenstrual age at discharge home. trials will be questioned andthe source of heterogen- 12. Corrected age-adequate performance (both compos- eity explored. ite cognitive score [Mental Developmental Index] Binary outcomes will be analysed using log binomial and composite motor score [Psychomotor Develop- regression models and results will be presented as risk mental Index] continuously and <85) on the two- ratios with 95% CI and associated two-sided P values. year Bayley scales of infant and toddler Continuous outcomes will be analysed using linear re- development-III; cerebral palsy rate including sever- gression models and results will be presented as differ- ity scaling with Gross Motor Function Classification ences in means with 95% CI and two-sided P values. Scale and a Child Behaviour Check List Any differences in treatment effect between pre-specified -questionnaire. subgroups of clinical characteristics at inclusion or at any 13. Fetal growth velocity, measured by expected/ time (gestational age, absent-reversed end-diastolic flow in observed average daily increase in ultrasound- the umbilical artery, placental growth factor maternal estimated serum levels, maternal disease present or absent) will be abdominal circumference. assessed by testing a treatment by subgroup interaction 14. Birth weight (grams, centiles). term within the model. 15. Gestational age at delivery (days). 16. Co-incidence and severity of the maternal syndrome Trial sequential analysis of pre-eclampsia/hemolysis, elevated liver enzymes, Trial sequential analysis will be applied to the meta- low platelets syndrome. analyses combining IPD and aggregate data because cu- 17. Maternal hypertension requiring treatment. mulative meta-analyses are at risk of producing random errors due to sparse data and repetitive testing of the ac- Safety monitoring (mother) cumulating data [16-21]. The underlying assumption of trial sequential analysis is that testing for significance 18. Serious adverse events. may be performed each time a new trial is added to the 19. Suspected unexpected serious adverse reaction. meta-analysis resulting in an increased risk of random 20. Anything else considered an adverse event. errors. We will add the trials according to the year of publication, and if more than one trial is published in a Safety monitoring (fetus) year, the trials will be added alphabetically according to the last name of the first author. On the basis of the 21. Serious adverse events. diversity-adjusted required information size, trial sequen- 22. Suspected unexpected serious adverse reaction. tial monitoring boundaries will be constructed. These 23. Anything else considered an adverse event. boundaries determine the statistical inference one may draw regarding the cumulative meta-analysis that has not Statistical analyses reached the required information size; if the trial sequen- The detailed statistical analysis plan will be prepared and tial monitoring boundary is crossed before the required agreed upon by the STRIDER IPD Study Group prior to information size is reached, firm evidence may perhaps be Ganzevoort et al. Systematic Reviews 2014, 3:23 Page 7 of 9 http://www.systematicreviewsjournal.com/content/3/1/23 established and further trials may turn out to be superflu- primarily on the primary outcome of the fetus and will ous. On the other hand, if the boundary is not surpassed, consider any adverse effects on the mother and/or child. it will most probably be necessary to continue conducting trials in order to detect or reject a certain intervention Ethics and management issues effect. Participants in the individual trials have given informed We will apply trial sequential analysis using a diversity- consent to participate in their respective trial. The data adjusted required information size calculated from an for this project are to be used for the purpose for which alpha error of 0.05, a beta error of 0.20, the control event they were originally collected and are available through proportion obtained from the results of the meta-analysis, an agreement between all trialists of the collaborative and a relative risk reduction of 20% for binary outcomes group. These trialists remain the custodians of their ori- with two or more trials to determine whether more trials ginal individual trial data at all times. are necessary on this topic (if the trial sequential alpha- spending monitoring boundary or the futility zone is Project management crossed, then more trials may be unnecessary). For For the purpose of this project, the international continuous outcomes, the required sample size will be STRIDER IPD Study Group will manage the project. calculated from an alpha error of 0.05, a beta error of Additionally, the coordinating investigator of each trial 0.20, the variance estimated from the meta-analysis re- will be invited to become a member of the STRIDER sults, and a minimal clinically relevant difference will IPD Study Group. The Study Group and statisticians be defined for each of the continuous outcomes. from the data management centre will develop the stat- istical analysis plan. The data management centre will be responsible for the storage and analyses of the IPD pro- Planned subgroup analyses ject data. We plan the following subgroup analyses: The Study Group meetings IPD trials compared to trials only providing Study Group face-to-face meetings will be organised at aggregate data. least twice during the study. Representatives of all eli- Trials with low risk of bias compared to trials with gible trials will be invited to attend those meetings. The high risk of bias. meetings will be scheduled, if possible, in conjunction Trials with dose of intervention below the median with international conferences. During those meetings, compared to trials with dose of intervention above various aspects of the project will be discussed with all the median. the collaborators, such as the project design and con- duct, the analysis plan, and the interpretation and Planned sensitivity analyses reporting of the results. To assess whether the results are robust to trial design The final results of the study will be presented to the and quality, sensitivity analyses will be performed on the collaborators for discussion. The main manuscript will primary outcome to test changes of the results after ex- be prepared by the STRIDER IPD Study Group. The re- clusion of trials with smaller sample size, centres with vised draft paper will be circulated for final comment smaller sample size, and trials/centres with high rate of and agreement prior to publication. Publications arising missing values. We will impute data for binary missing from these data will be authored by all members of the outcomes using various scenarios, namely best-best ana- STRIDER IPD Study Group on behalf of all other collab- lysis, worst-worst analysis, best-worst analysis, and worst- orators participating, who will be acknowledged within best analysis. the manuscript. Multiple comparisons Discussion A very large number of outcomes are being investigated Severe early-onset FGR due to placental insufficiency is in this study, which increases the risk of observing ‘false associated with a high risk of perinatal morbidity with positive’ results due to multiplicity. However, all out- long-lasting sequelae and mortality. Placental insuffi- comes are important in giving a full clinical picture that ciency is the result of abnormal formation and function considers the benefits and risks to both mothers and in- of the placenta (placentation) with inadequate remodel- fants. We do not plan formal statistical adjustment of P ling of the maternal spiral (uteroplacental) arteries. values to account for multiple comparisons due to the There is currently no therapy available with demon- non-independence of outcomes in this study. Results strated effectiveness, making monitoring and timely de- will be interpreted with caution. Decisions as to whether livery of the child the only treatment option. Both the intervention should be recommended will be based in vitro and in vivo evidence suggests sildenafil citrate as Ganzevoort et al. Systematic Reviews 2014, 3:23 Page 8 of 9 http://www.systematicreviewsjournal.com/content/3/1/23 a therapeutic strategy to improve uteroplacental blood collaborations, and reviewed the manuscript. All authors read and approved the final manuscript. flow, fetal growth, and meaningful outcomes in FGR pregnancies. Acknowledgements On the basis of preliminary research, some centres are New Zealand/Australia: funded by the Health Research Council (New already adopting the treatment with sildenafil. There is, Zealand). UK/Ireland study: funded by the Medical Research Council (United Kingdom). however, significant uncertainty regarding the true health benefits, if any. Moreover, potential harm is not yet ex- Author details cluded. Fetotoxicity has not been described yet, but is the- Department of Obstetrics and Gynecology, Academic Medical Centre, Room: H4-278, PO BOX 22660, 1100 DD Amsterdam, The Netherlands. oretically possible. For these reasons, evidence from Department of Women’s and Children’s Health, Liverpool Women’s Hospital, randomised clinical trials is needed before widespread im- University of Liverpool, Crown Street, Liverpool L8 7SS, UK. Department of plementation [31]. Obstetrics and Gynaecology, University of British Colombia, Rm V3-339, 950 West 28th Avenue, Vancouver, BC V5Z 4H4, Canada. The Irish Centre for Given the sample size issues, no single randomised Fetal and Neonatal Translational Research, University College Cork, Cork, clinical trial will ultimately be likely to give the answers Ireland. Department of Obstetrics and Fetal Medicine, St George’s Hospital, to the questions posed in this study. For this reason, Blackshaw Road, London SW17 0QT, UK. Department of Neonatology, Academic Medical Centre, PO BOX 22660, 1100 DD Amsterdam, The simultaneous smaller and medium-sized randomised Netherlands. Department 7812, Copenhagen Trial Unit, Centre for Clinical clinical trials have been planned and funded. The present Intervention Research, Rigshospitalet, Copenhagen University Hospital, pre-planned systematic review with IPD meta-analysis will Blegdamsvej 9, DK 2100 Copenhagen, Denmark. Department of Obstetrics and Gynaecology, Academic Medical Centre, Room H4-213, PO Box 22700, allow more definite answers, including effect modifica- 1105 DE Amsterdam, The Netherlands. Liggins Institute, University of tions in relevant subgroups. Furthermore, the totality of Auckland, Level 2, Building 505, 85 Park Rd, Grafton, Private Bag 92019, evidence from IPD trials and trials reporting only aggre- Auckland 1142, New Zealand. gate data can be assessed and evaluated with GRADE Received: 30 October 2013 Accepted: 27 February 2014 methodology [32]. All weighing of results should take into Published: 11 March 2014 consideration risks of bias and aim to minimise these: sys- tematic errors (bias) due to methodology and design; risks References 1. Lee MJ, Conner EL, Charafeddine L, Woods JR Jr, Del Priore G: A critical of academic bias if there are conflicting results between birth weight and other determinants of survival for infants with severe studies from this GONet group and other studies; risks of intrauterine growth restriction. Ann N Y Acad Sci 2001, 943:326–339. random errors (play of chance) [33]. According to our 2. Ananth CV, Vintzileos AM: Maternal-fetal conditions necessitating a medical intervention resulting in preterm birth. Am J Obstet Gynecol 2006, knowledge, this is the first time a protocol for a systematic 195(6):1557–1563. review has been finalised and published before the first 3. Petersen SG, Wong SF, Urs P, Gray PH, Gardener GJ: Early onset, severe larger trial has been launched on a topic. With the All- fetal growth restriction with absent or reversed end-diastolic flow velocity waveform in the umbilical artery: perinatal and long-term Trials initiative and other initiatives for transparency in outcomes. Aust N Z J Obstet Gynaecol 2009, 49(1):45–51. clinical research, we hope that the number of trials provid- 4. Batton DG, DeWitte DB, Espinosa R, Swails TL: The impact of fetal ing IPD for this systematic review will be large. compromise on outcome at the border of viability. Am J Obstet Gynecol 1998, 178(5):909–915. 5. Wood NS, Marlow N, Costeloe K, Gibson AT, Wilkinson AR: Neurologic and Abbreviations developmental disability after extremely preterm birth. EPICure Study FGR: Fetal growth restriction; GONet: Global obstetrics network; Group. N Engl J Med 2000, 343(6):378–384. IPD: Individual participant data; NICU: Neonatal intensive care unit; 6. Larroque B, Ancel PY, Marchand-Martin L, Cambonie G, Fresson J, Pierrat V, STRIDER: Sildenafil therapy in dismal prognosis early-onset intrauterine Rozé JC, Marpeau L, Thiriez G, Alberge C, Bréart G, Kaminski M, Marret S, growth restriction; UBC: University of British Columbia. Epipage Study group: Special care and school difficulties in 8-year-old very preterm children: the Epipage cohort study. PLoS One 2011, 6(7):e21361. Competing interests 7. Ganzevoort W, Rep A, De Vries JI, Bonsel GJ, Wolf H, PETRA-investigators: None of the authors have competing interests to report. Authors PB, ZA, LK, Prediction of maternal complications and adverse infant outcome at AP, WG, BM, AvW, and PvD are involved in successful/ongoing funding admission for temporizing management of early-onset severe hypertensive applications for the studies that form the basis of this IPD (Health Research disorders of pregnancy. Am J Obstet Gynecol 2006, 195(2):495–503. Council, New Zealand; UK Medical Research Council, United Kingdom; 8. Guellec I, Lapillonne A, Renolleau S, Charlaluk ML, Roze JC, Marret S, Vieux R, ZonMW, The Netherlands; and CIHR, Canada). Monique K, Ancel PY, EPIPAGE Study Group: Neurologic outcomes at school age in very preterm infants born with severe or mild growth Authors’ contributions restriction. Pediatrics 2011, 127(4):e883–e891. WG participated in the study design and drafted the manuscript. ZA 9. Walther FJ, Den Ouden AL, Verloove-Vanhorick SP: Looking back in time: participated in the design of the study and reviewed the manuscript. PvD outcome of a national cohort of very preterm infants born in The developed the hypotheses, participated in the design of the study, instigated Netherlands in 1983. Early Hum Dev 2000, 59(3):175–191. the international collaborations, and reviewed the manuscript. LK developed 10. Stanley JL, Andersson IJ, Poudel R, Rueda-Clausen CF, Sibley CP, Davidge ST, the hypotheses, participated in the design of the study, instigated the Baker PN: Sildenafil citrate rescues fetal growth in the catechol-O-methyl international collaborations, and reviewed the manuscript. AP participated in transferase knockout mouse model. Hypertension 2012, 59(5):1021–1028. the design of the study and reviewed the manuscript. AvW participated in 11. Wareing M, Myers JE, O’Hara M, Baker PN: Sildenafil citrate (Viagra) the design of the study particularly neonatal outcome analysis and reviewed enhances vasodilatation in fetal growth restriction. J Clin Endocrinol the manuscript. CG participated in the design of the study, particularly the Metab 2005, 90(5):2550–2555. statistics section and reviewed the manuscript. BM participated in the design 12. Herraiz S, Pellicer B, Serra V, Cauli O, Cortijo J, Felipo V, Pellicer A: Sildenafil of the study and reviewed the manuscript. PB developed the hypotheses, citrate improves perinatal outcome in fetuses from pre-eclamptic rats. participated in the design of the study, instigated the international BJOG 2012, 119(11):1394–1402. Ganzevoort et al. Systematic Reviews 2014, 3:23 Page 9 of 9 http://www.systematicreviewsjournal.com/content/3/1/23 13. Samangaya RA, Mires G, Shennan A, Skillern L, Howe D, McLeod A, Baker PN: A randomised, double-blinded, placebo-controlled study of the phosphodiesterase type 5 inhibitor sildenafil for the treatment of preeclampsia. Hypertens Pregnancy 2009, 28(4):369–382. 14. von Dadelszen P, Dwinnell S, Magee LA, Carleton BC, Gruslin A, Lee B, Lim KI, Liston RM, Miller SP, Rurak D, Sherlock RL, Skoll MA, Wareing MM, Baker PN, Research into Advanced Fetal Diagnosis and Therapy (RAFT) Group: Sildenafil citrate therapy for severe early-onset intrauterine growth restriction. BJOG 2011, 118(5):624–628. 15. GONet. Available from: http://www.globalobstetricsnetwork.org. 16. Broeze KA, Opmeer BC, van der Veen F, Bossuyt PM, Bhattacharya S, Mol BW: Individual patient data meta-analysis: a promising approach for evidence synthesis in reproductive medicine. Hum Reprod Update 2010, 16(6):561–567. 17. Wetterslev J, Thorlund K, Brok J, Gluud C: Trial sequential analysis may establish when firm evidence is reached in cumulative meta-analysis. J Clin Epidemiol 2008, 61(1):64–75. 18. Thorlund K, Devereaux PJ, Wetterslev J, Guyatt G, Ioannidis JP, Thabane L, Gluud LL, Als-Nielsen B, Gluud C: Can trial sequential monitoring boundaries reduce spurious inferences from meta-analyses? Int J Epidemiol 2009, 38 (1):276–286. 19. Thorlund K, Imberger G, Walsh M, Chu R, Gluud C, Wetterslev J, Guyatt G, Devereaux PJ, Thabane L: The number of patients and events required to limit the risk of overestimation of intervention effects in meta-analysis–a simulation study. PLoS One 2011, 6(10):e25491. 20. Wetterslev J, Thorlund K, Brok J, Gluud C: Estimating required information size by quantifying diversity in random-effects model meta-analyses. BMC Med Res Methodol 2009, 9:86. 21. User Manual for Trial Sequential Analysis (TSA) from the Copenhagen Trial Unit, Center for Clinical Intervention Research, Rigshospitalet, Copenhagen, Denmark. http://www.ctu.dk/tsa/. 22. WHO Trial Portal. http://apps.who.int/trialsearch/. 23. Current Controlled Trials. http://www.controlled-trials.com. 24. The Australian and New Zealand Trials Register. http://www.anzctr.org.au. 25. ISRCTN Register. http://www.isrctn.org. 26. The Cochrane Handbook for Systematic Reviews of Interventions. http://www. cochrane.org/training/cochrane-handbook. 27. Savovic J, Jones H, Altman D, Harris R, Juni P, Pildal J, Als-Nielsen B, Balk E, Gluud C, Gluud L, Ioannidis J, Schulz K, Beynon R, Welton N, Wood L, Moher D, Deeks J, Sterne J: Influence of reported study design characteristics on intervention effect estimates from randomised controlled trials: combined analysis of meta-epidemiological studies. Health Technol As- sess 2012, 16(35):1–82. 28. Savovic J, Jones HE, Altman DG, Harris RJ, Juni P, Pildal J, Als-Nielsen B, Balk EM, Gluud C, Gluud LL, Ioannidis JP, Schulz KF, Beynon R, Welton NJ, Wood L, Moher D,Deeks JJ,SterneJA: Influence of reported study design characteristics on intervention effect estimates from randomized, controlled trials. AnnInternMed 2012, 157(6):429–438. 29. Wood L, Egger M, Gluud LL, Schulz KF, Juni P, Altman DG, Gluud C, Martin RM, Wood AJ, Sterne JA: Empirical evidence of bias in treatment effect estimates in controlled trials with different interventions and outcomes: meta-epidemiological study. BMJ 2008, 336(7644):601–605. 30. Lundh A, Sismondo S, Lexchin J, Busuioc OA, Bero L: Industry sponsorship and research outcome. Cochrane Database Syst Rev 2012, 12, MR000033. 31. Jakobsen JCGC: The necessity of randomized clinical trials. Br J Med Med Res 2013, 3(4):16. 32. Grade Working Group. http://www.gradeworkinggroup.org. 33. Keus F, Wetterslev J, Gluud C, van Laarhoven CJ: Evidence at a glance: error matrix approach for overviewing available evidence. BMC Med Res Submit your next manuscript to BioMed Central Methodol 2010, 10:90. and take full advantage of: doi:10.1186/2046-4053-3-23 Cite this article as: Ganzevoort et al.: STRIDER: Sildenafil therapy in • Convenient online submission dismal prognosis early-onset intrauterine growth restriction – a protocol • Thorough peer review for a systematic review with individual participant data and aggregate • No space constraints or color figure charges data meta-analysis and trial sequential analysis. Systematic Reviews 2014 3:23. • Immediate publication on acceptance • Inclusion in PubMed, CAS, Scopus and Google Scholar • Research which is freely available for redistribution Submit your manuscript at www.biomedcentral.com/submit http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Systematic Reviews Springer Journals

STRIDER: Sildenafil therapy in dismal prognosis early-onset intrauterine growth restriction – a protocol for a systematic review with individual participant data and aggregate data meta-analysis and trial sequential analysis

Download PDF
9 pages

Loading next page...
 
/lp/springer-journals/strider-sildenafil-therapy-in-dismal-prognosis-early-onset-qYkku00bDW

References (68)

Publisher
Springer Journals
Copyright
Copyright © 2014 by Ganzevoort et al.; licensee BioMed Central Ltd.
Subject
Medicine & Public Health; Medicine/Public Health, general; Biomedicine general; Statistics for Life Sciences, Medicine, Health Sciences
eISSN
2046-4053
DOI
10.1186/2046-4053-3-23
pmid
24618418
Publisher site
See Article on Publisher Site

Abstract

Background: In pregnancies complicated by early-onset extreme fetal growth restriction, there is a high risk of preterm birth and an overall dismal fetal prognosis. Sildenafil has been suggested to improve this prognosis. The first aim of this review is to assess whether sildenafil benefits or harms these babies. The second aim is to analyse if these effects are modified in a clinically meaningful way by factors related to the women or the trial protocol. Methods/Design: The STRIDER (Sildenafil Therapy In Dismal prognosis Early-onset intrauterine growth Restriction) Individual Participant Data (IPD) Study Group will conduct a prospective IPD and aggregate data systematic review with meta-analysis and trial sequential analysis. The STRIDER IPD Study Group started trial planning and funding applications in 2012. Three trials will be launched in 2014, recruiting for three years. Further trials are planned to commence in 2015. The primary outcome for babies is being alive at term gestation without evidence of serious adverse neonatal outcome. The latter is defined as severe central nervous system injury (severe intraventricular haemorrhage (grade 3 and 4) or cystic periventricular leukomalacia, demonstrated by ultrasound and/or magnetic resonance imaging) or other severe morbidity (bronchopulmonary dysplasia, retinopathy of prematurity requiring treatment, or necrotising enterocolitis requiring surgery). The secondary outcomes are improved fetal growth velocity assessed by ultrasound abdominal circumference measurements, gestational age and birth weight (centile) at delivery, and age-adequate performance on the two-year Bayley scales of infant and toddler development-III (composite cognitive score and composite motor score). Subgroup and sensitivity analyses in the IPD meta-analysis include assessment of the influence of several patient characteristics: an abnormal or normal serum level of placental growth factor, absent/reversed umbilical arterial end diastolic flow at commencement of treatment, and other patient characteristics available at baseline such as gestational age and estimated fetal weight. The secondary outcomes for mothers include co-incidence and severity of the maternal syndrome of pre-eclampsia, mortality, and other serious adverse events. Discussion: Trials are expected to start in 2013–2014 and end in 2016–2017. Data analyses of individual trials are expected to finish in 2019. Given the pre-planned and agreed IPD protocol, these results should be available in 2020. Keywords: Adverse neonatal outcome, Fetal growth restriction, Individual participant data meta-analysis, Sildenafil * Correspondence: [email protected] Department of Obstetrics and Gynecology, Academic Medical Centre, Room: H4-278, PO BOX 22660, 1100 DD Amsterdam, The Netherlands Full list of author information is available at the end of the article © 2014 Ganzevoort et al.; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Ganzevoort et al. Systematic Reviews 2014, 3:23 Page 2 of 9 http://www.systematicreviewsjournal.com/content/3/1/23 Background the belief that rest will enhance the uteroplacental circula- Severe early-onset fetal growth restriction (FGR) compli- tion at the expense of that to the glutei and quadriceps cates approximately 0.4% of pregnancies and severely in- muscles. creases the risk of perinatal morbidity and mortality. This There is evidence from ex vivo and animal models of is particularly due to premature delivery, both for fetal growth restriction that the phosphodiesterase 5 inhibitor and for secondary maternal indications such as the devel- sildenafil citrate increases average pup birth weight and opment of pre-eclampsia [1]. Placental disease, conse- improves uteroplacental blood flow (umbilical artery, quent on deficient uteroplacental blood flow, includes uterine artery) [10-12]. Sildenafil citrate may thus offer a FGR, pre-eclampsia, and placental abruption and has been potential therapeutic strategy to improve uteroplacental implicated in more than 50% of iatrogenic premature blood flow in human pregnancies complicated by severe births [2]. For this reason, the problem of severe FGR FGR [12]. There is evidence, albeit limited, to support forms a substantial portion of the population that tertiary the use of sildenafil in human pregnancy. A small trial in care centres care for. women with early-onset pre-eclampsia (a maternal pres- The effect of severe early-onset FGR is particularly sig- entation of placental compromise) showed no resolution nificant: of those born alive, less than a third will survive of pre-eclampsia, but an effect on birth weight [13]. In their neonatal intensive care unit (NICU) stay without addition, a small cohort study showed a tendency to- significant neurodevelopmental sequelae. Survival rates wards more live-born children that survived intact to pri- for severely growth-restricted fetuses very remote from mary discharge after sildenafil intervention [14]. From all term (<28 weeks’ gestation) vary from 7% to 33% [1,3,4]. observations thus far, there are no indications that sildena- As these early-onset FGR children are born very pre- fil in the second trimester has significant fetotoxicity. term, there are significant risks of neonatal mortality, Coordinated initiatives have been launched and ap- major and minor morbidity, and long-term health seque- proved/shortlisted for funding in networks in the United lae [5,6]. These are not only strongly gestational-age re- Kingdom/Ireland, The Netherlands, and New Zealand/ lated [7], but are also related to FGR [8]. Although most Australia. Initiatives in Canada and the United States of of the surviving children are without severe disabilities America are in an earlier phase. These trials are embed- at school age, many of them meet serious difficulties in ded in a global network (Global Obstetrics Network – everyday life and the burden of mild developmental ab- GONet) [15]. This network was initiated to provide a normalities and behavioural and learning disorders in- forum for international interaction and foster communi- creases with age. In adolescents, as many as 40% of the cation between groups that perform clinical studies in survivors will not be able to become fully independent obstetrics. GONet has started the collaborative initiative adults [9]. to resolve the underlying question: does sildenafil versus The direct costs of care include the increased cost of placebo improve outcomes in early-onset severely growth intensive antenatal surveillance (with or without a period restricted fetuses with a dismal prognosis? Within each of hospitalisation), caesarean delivery (if considered vi- part of the network, stand-alone randomised clinical trials able), NICU care, routine post-NICU follow-up, need for will be conducted; a prospective systematic review with special schooling and educational support (speech ther- meta-analyses and trial sequential analyses of individual apy, physical therapy, psychological support), and, often, participant data (IPD) and aggregate data will then be per- specialised neurodevelopmental assessments and inter- formed. The concerted effort in the design and conduct of ventions. Such costs are in addition to the indirect costs these large international studies provides an efficient use to families (grief, lost wages, supportive services, and re- of scarce resources and will optimise resolution of the re- lationship stress) and represent a considerable economic search questions, in addition to facilitating adequate burden. Safe pregnancy prolongation accompanied by power for subgroup analyses. additional fetal growth would be associated with signifi- cant societal and financial savings. Methods/Design The use of ultrasound Doppler waveform analysis in Study design pregnancies complicated by FGR suggests compromised Prospectively planned systematic review with IPD and uteroplacental circulation and placental hypoperfusion. aggregate data meta-analysis [16] and trial sequential Currently there are no specific evidence-based therapies analysis [17-21] of randomised clinical trials comparing for placental insufficiency and early-onset severe FGR. sildenafil with matching placebo tablets or no interven- Non-specific interventions include primarily lifestyle mod- tion for treatment of severe early-onset FGR. ifications, such as reducing or stopping work, stopping aerobic exercise, rest at home, and hospital admission for Inclusion and exclusion criteria for the studies rest and surveillance. These interventions, which are not Studies, published or unpublished, will be included if supported by evidence from randomised trials, are used in they are randomised clinical trials. Other study designs Ganzevoort et al. Systematic Reviews 2014, 3:23 Page 3 of 9 http://www.systematicreviewsjournal.com/content/3/1/23 for assessment of benefits will not be included. For the join the STRIDER Study Group have agreed a common assessment of harm we will also include observational core data set and have agreed to supply their trial’s IPD studies. Given the pre-planned nature of the IPD meta- to a central repository. Randomisation will be centrally analysis, most or all studies will have been identified be- controlled using an on-line computerised randomisation fore the start of the individual trials and will participate service. All have agreed to collect all relevant variables in central data collection. If other trials, yet unknown, on baseline characteristics and outcomes (the primary are conducted simultaneously, their eligibility will be and secondary outcomes of this protocol). All have assessed independently and unblinded for author and agreed to centrally collect data. The University of British journal by two members of the Sildenafil Therapy In Columbia (UBC) will host data management through Dismal prognosis Early-onset intrauterine growth Re- the UBC Perinatal Clinical Trials Unit. Wherever pos- striction (STRIDER) IPD Study Group. Any differences sible, data will be collected as continuous rather than in opinion regarding eligibility will be resolved by dis- dichotomous measurements. All trialists have agreed cussion. If the study design is suboptimal for IPD meta- to collect information on the number of eligible but analysis or the authors do not cooperate, the study will not-included patients. If the patient consents, a number be included for aggregate data analysis. of relevant baseline and outcome parameters will be We will use the Cochrane Central Register of Con- collected. trolled Trials (CENTRAL), MEDLINE, Embase, LILACS, If other relevant randomised clinical trials are identi- and Science Citation Index Expanded using the terms fied, data collected from those studies will include all “fetal growth retardation” and “sildenafil”. In addition, women randomised and coded for anonymity (date of we will access the WHO Trial Portal [22], Current Con- birth, centre identification), baseline data for descriptive trolled Trials [23], the Australian and New Zealand Trials purposes and analyses, details of the intervention given Register [24], and the ISRCTN register [25] to identify re- (date of randomisation, allocated intervention), and out- cently completed or ongoing trials. Experts in the field comes, to allow planned analyses. Trialists will provide and trialists will be asked if they know of any unpublished non-identified IPD in any convenient format by encrypted, or other trials. The currently known trials that will be con- electronic transfer where possible or other means as ducted and included are: needed. The individual trial data will be recoded as re- quired, and stored in the secure database that will only be UK/Ireland: Planned sample size 112; primary accessible by authorised personnel of the UBC Perinatal outcome: prolongation of gestational age at birth. Clinical Trials Unit. Trialists will be asked to verify all The Netherlands: Planned sample size 354; primary recoded data prior to any analysis and the data will not be outcome: infant survival to term age without evidence used for any other purpose without permission of all of serious adverse neonatal outcome. collaborators. New Zealand/Australia: Planned sample size 122; The data will be checked with respect to range, in- primary outcome: fetal growth velocity. ternal consistency, missing or extreme values, errors, Canada: Planned sample size 90; primary outcome: and consistency with published reports. Trial details, fetal growth velocity. such as randomisation methods, blinding, and interven- tion details, will be crosschecked against published reports, Types of participants trial protocols, and data collection sheets. Inconsistencies Women with a singleton pregnancy between 20 + 0 and or missing data will be discussed with the individual trial- 29 + 6 weeks with severe fetal growth restriction of likely ists and attempts will be made to resolve any problems by placental origin, and with estimated significant likeli- consensus. hood of fetal/neonatal death. Data items to be collected Type of intervention Trial level information Sildenafil versus placebo tablet orally or versus no inter- vention until fetal death, delivery, or 32 weeks of gesta- 1. Dates the trial opened and closed accrual. tion (whichever comes first). Co-interventions will be 2. Number of participants randomised. allowed provided they are administered similarly in the 3. Informed consent procedures. intervention groups. 4. Methods of random allocation. 5. Stratification factors (if any). Data collection and management 6. Methods of allocation concealment. The investigators of each individual trial have designed 7. Blinding procedures for outcome assessment. each trial together before launch and patient recruit- 8. Details of the planned intervention in the drug arm: ment. All of the listed eligible trials that have agreed to frequency, timing, doses. Ganzevoort et al. Systematic Reviews 2014, 3:23 Page 4 of 9 http://www.systematicreviewsjournal.com/content/3/1/23 9. Details of the planned intervention in the placebo 2. Date and time of delivery. arm: frequency, timing, doses. 3. Birth weight, 5 minute Apgar scores, sex. 4. Mode of delivery. Participant-level information – maternal characteristics at 5. Neonatal morbidity. study entry 6. Date and time if neonatal death. 7. Childhood follow-up assessment. 1. Unique identification coded for anonymity. 8. Adverse events and adverse effects for the neonate. 2. Maternal age. 3. Body mass index. Participant-level information – data on actual trial intervention 4. Parity. 5. Abdominal circumference. 1. Type of treatment given (placebo, sildenafil, dose). 6. Ethnicity. 2. Number of doses actually given. 7. Hypertensive disease (type of). 3. Interval between first dose and 32 weeks or delivery 8. Uterine artery notching/pulsatility index. or fetal death. 9. Placental growth factor levels in maternal serum. 4. Total actual drug exposure. 10.Treatment (e.g., aspirin, low molecular weight heparin, oral steroids, insulin). Bias risk assessments This project will use the Cochrane Collaboration methods Participant-level information – fetal characteristics at study entry for assessing risk of bias within individual trials [26]: 1. Estimated due date. Allocation sequence generation 2. Date and time of randomisation. 3. Estimated fetal weight, abdominal circumference. Low risk of bias: sequence generation was achieved 4. Absent or reversed end-diastolic flow of the umbil- using computer random number generation or a ical artery, pulsatility index. random number table. Drawing lots, tossing a coin, 5. Pulsatility index of middle cerebral artery. shuffling cards, and throwing dice are adequate if performed by an independent person not otherwise Participant-level information – maternal information after involved in the trial. trial entry Uncertain risk of bias: the method of sequence generation was not specified. 1. Date and time of prenatal corticosteroid treatment. High risk of bias: the sequence generation method 2. Hypertensive disease (type of). was not random. 3. Highest blood pressure measured. 4. Proteinuria (maximum level recorded). Allocation concealment 5. Platelet count (lowest recorded). 6. Adverse events and adverse effects for the woman at Low risk of bias: the participant allocations could not time of treatment and post-partum. have been foreseen in advance of, or during, enrolment. Allocation was controlled by a central and independent Participant-level information – fetal information after trial entry randomisation unit. The allocation sequence was unknown to the investigators (for example, if the 1. Sequential assessments of the umbilical artery: absent allocation sequence was hidden in sequentially or reversed end-diastolic flow, pulsatility index. numbered, opaque, and sealed envelopes). 2. Sequential assessments of estimated fetal weight, Uncertain risk of bias: the method used to conceal the abdominal circumference. allocation was not described so that intervention 3. Sequential assessments of the middle cerebral artery: allocations may have been foreseen in advance of, or pulsatility index. during, enrolment. 4. Estimated date and time if fetal death. High risk of bias: the allocation sequence was likely to be 5. Adverse events and adverse effects for the fetus. known to the investigators who assigned the participants. Participant-level information – neonatal information after Blinding of participants, personnel, and outcome assessors trial entry Low risk of bias: blinding was performed adequately or 1. Unique baby information and mother identification the assessment of outcomes was not likely to be coded for anonymity. influenced by lack of blinding. Ganzevoort et al. Systematic Reviews 2014, 3:23 Page 5 of 9 http://www.systematicreviewsjournal.com/content/3/1/23 Uncertain risk of bias: there was insufficient High risk of bias: the trial is sponsored by the information to assess whether blinding was likely to industry. induce bias on the results. High risk of bias: no blinding or incomplete blinding, Other bias and the assessment of outcomes were likely to be influenced by lack of blinding. Low risk of bias: the trial appears to be free of other components that could put it at risk of bias. Incomplete outcome data Uncertain risk of bias: the trial may or may not be free of other components that could put it at risk of Low risk of bias: missing data were unlikely to make bias. treatment effects depart from plausible values. High risk of bias: there are other factors in the trial Sufficient methods, such as multiple imputation, have that could put it at risk of bias. been employed to handle missing data. Trials assessed as having ‘low risk of bias’ in all of the Uncertain risk of bias: there was insufficient individual domains specified in the review were information to assess whether missing data in considered ‘trials with low risk of bias’. Trials assessed combination with the method used to handle missing as having ‘uncertain risk of bias’ or ‘high risk of bias’ data were likely to induce bias on the results. in one or more of the specified in the review High risk of bias: the results were likely to be biased individual domains were considered trials with ‘high due to missing data. risk of bias’ [26-30]. Selective outcome reporting Dealing with missing data We intend to perform an intention-to-treat analysis when- Low risk of bias: all outcomes were pre-defined and re- ever possible. We will impute data for binary outcomes ported or all clinically relevant and reasonably expected using various scenarios, namely best-best analysis, worst- outcomes were reported. worst analysis, best-worst analysis, and worst-best analysis. Uncertain risk of bias: it was unclear whether all pre- For continuous outcomes, we will use available-participant defined and clinically relevant and reasonably ex- analysis. We will impute the standard deviation from P pected outcomes were reported. values according to the instructions given in the Cochrane High risk of bias: one or more clinically relevant and Handbook for Systematic Reviews of Intervention [26] and reasonably expected outcomes were not reported, and use the median for the meta-analysis when the mean is not data on these outcomes were likely to have been available. If it is not possible to calculate the standard devi- recorded. ation from the P value or the confidence intervals (CIs), we For a trial to be assessed with low risk of bias in the will impute the standard deviation as the highest standard selective outcome reporting domain, the trial should deviation in the other trials included under that outcome, have been registered either on the www.clinicaltrials. fully recognising that this form of imputation will decrease gov website or a similar register, or there should be a the weight of the trial for calculation of mean differences protocol, e.g., published in a paper journal. In the case and bias the effect estimate to no effect in the case of stan- when the trial was run and published in the years dardised mean difference. when trial registration was not required, we carefully scrutinised all publications reporting on the trial to Outcomes identify the trial objectives and outcomes. If usable The primary outcome is defined as: data on all outcomes specified in the trial objectives Neonatal survival until term age without serious neo- were provided in the publications results section, then natal morbidity defined as secondary outcomes 3–6. The the trial was considered low risk of bias trial in the outcome needs to be assessed formally at term age by Selective outcome reporting domain. clinic visit (or as a last-resort alternative by telephone). Secondary outcomes are: Industry bias 1. Any perinatal deaths (any know death irrespective of Low risk of bias: the trial appears to be free of industry cause). sponsorship that may manipulate the trial design, 2. Perinatal deaths in normally formed infants (absence conductance, or results of the trial. of severe congenital malformation, infection, or Uncertain risk of bias: the trial may or may not be genetic abnormality). free of for-profit bias as no information on clinical 3. Severe central nervous system injury (diagnosed by trial support or sponsorship is provided. ultrasound and/or magnetic resonance imaging) Ganzevoort et al. Systematic Reviews 2014, 3:23 Page 6 of 9 http://www.systematicreviewsjournal.com/content/3/1/23 periventricular leukomalacia grade II or more, or the data analyses. Any analyses conducted will be based intracerebral haemorrhage grade III or more, or on the checked and updated individual participant data hydrocephalus. from all available trials. All randomised participants with 4. Bronchopulmonary dysplasia (the requirement of outcome data available will be included in the analyses, ambulatory oxygen therapy >36 weeks corrected which will be performed on an intention-to-treat basis, gestational age). according to the treatment allocation at randomisation. 5. Retinopathy of prematurity (requiring treatment For each of the outcomes (as well as the individual such as laser therapy; grade 2/3 or more). components of each composite outcome), a one-stage 6. Necrotising enterocolitis (requiring surgery). approach to analysis will be taken so that the IPD from 7. Need for inotropes or vasopressors: number of days all eligible trials are included in a single model. Fitting a on inotropes or vasopressors, days on one, two, or single model for each outcome variable will enable the three different inotropes or vasopressors. variation across trials to be accounted for within the 8. Patent ductus arteriosus needing medical or surgical model by including a fixed trial effect. A treatment by treatment. trial interaction term will be tested to assess heterogeneity 9. Number of septic episodes, defined as culture of treatment effect across trials. If excessive statistical het- proven or clinical, with need for antibiotics for 5 or erogeneity in treatment effect or inconsistency across tri- more days. als is detected (that is, if the trial by treatment interaction 10. Total number of days on artificial ventilation. term is significant), then the rationale for combining 11. Postmenstrual age at discharge home. trials will be questioned andthe source of heterogen- 12. Corrected age-adequate performance (both compos- eity explored. ite cognitive score [Mental Developmental Index] Binary outcomes will be analysed using log binomial and composite motor score [Psychomotor Develop- regression models and results will be presented as risk mental Index] continuously and <85) on the two- ratios with 95% CI and associated two-sided P values. year Bayley scales of infant and toddler Continuous outcomes will be analysed using linear re- development-III; cerebral palsy rate including sever- gression models and results will be presented as differ- ity scaling with Gross Motor Function Classification ences in means with 95% CI and two-sided P values. Scale and a Child Behaviour Check List Any differences in treatment effect between pre-specified -questionnaire. subgroups of clinical characteristics at inclusion or at any 13. Fetal growth velocity, measured by expected/ time (gestational age, absent-reversed end-diastolic flow in observed average daily increase in ultrasound- the umbilical artery, placental growth factor maternal estimated serum levels, maternal disease present or absent) will be abdominal circumference. assessed by testing a treatment by subgroup interaction 14. Birth weight (grams, centiles). term within the model. 15. Gestational age at delivery (days). 16. Co-incidence and severity of the maternal syndrome Trial sequential analysis of pre-eclampsia/hemolysis, elevated liver enzymes, Trial sequential analysis will be applied to the meta- low platelets syndrome. analyses combining IPD and aggregate data because cu- 17. Maternal hypertension requiring treatment. mulative meta-analyses are at risk of producing random errors due to sparse data and repetitive testing of the ac- Safety monitoring (mother) cumulating data [16-21]. The underlying assumption of trial sequential analysis is that testing for significance 18. Serious adverse events. may be performed each time a new trial is added to the 19. Suspected unexpected serious adverse reaction. meta-analysis resulting in an increased risk of random 20. Anything else considered an adverse event. errors. We will add the trials according to the year of publication, and if more than one trial is published in a Safety monitoring (fetus) year, the trials will be added alphabetically according to the last name of the first author. On the basis of the 21. Serious adverse events. diversity-adjusted required information size, trial sequen- 22. Suspected unexpected serious adverse reaction. tial monitoring boundaries will be constructed. These 23. Anything else considered an adverse event. boundaries determine the statistical inference one may draw regarding the cumulative meta-analysis that has not Statistical analyses reached the required information size; if the trial sequen- The detailed statistical analysis plan will be prepared and tial monitoring boundary is crossed before the required agreed upon by the STRIDER IPD Study Group prior to information size is reached, firm evidence may perhaps be Ganzevoort et al. Systematic Reviews 2014, 3:23 Page 7 of 9 http://www.systematicreviewsjournal.com/content/3/1/23 established and further trials may turn out to be superflu- primarily on the primary outcome of the fetus and will ous. On the other hand, if the boundary is not surpassed, consider any adverse effects on the mother and/or child. it will most probably be necessary to continue conducting trials in order to detect or reject a certain intervention Ethics and management issues effect. Participants in the individual trials have given informed We will apply trial sequential analysis using a diversity- consent to participate in their respective trial. The data adjusted required information size calculated from an for this project are to be used for the purpose for which alpha error of 0.05, a beta error of 0.20, the control event they were originally collected and are available through proportion obtained from the results of the meta-analysis, an agreement between all trialists of the collaborative and a relative risk reduction of 20% for binary outcomes group. These trialists remain the custodians of their ori- with two or more trials to determine whether more trials ginal individual trial data at all times. are necessary on this topic (if the trial sequential alpha- spending monitoring boundary or the futility zone is Project management crossed, then more trials may be unnecessary). For For the purpose of this project, the international continuous outcomes, the required sample size will be STRIDER IPD Study Group will manage the project. calculated from an alpha error of 0.05, a beta error of Additionally, the coordinating investigator of each trial 0.20, the variance estimated from the meta-analysis re- will be invited to become a member of the STRIDER sults, and a minimal clinically relevant difference will IPD Study Group. The Study Group and statisticians be defined for each of the continuous outcomes. from the data management centre will develop the stat- istical analysis plan. The data management centre will be responsible for the storage and analyses of the IPD pro- Planned subgroup analyses ject data. We plan the following subgroup analyses: The Study Group meetings IPD trials compared to trials only providing Study Group face-to-face meetings will be organised at aggregate data. least twice during the study. Representatives of all eli- Trials with low risk of bias compared to trials with gible trials will be invited to attend those meetings. The high risk of bias. meetings will be scheduled, if possible, in conjunction Trials with dose of intervention below the median with international conferences. During those meetings, compared to trials with dose of intervention above various aspects of the project will be discussed with all the median. the collaborators, such as the project design and con- duct, the analysis plan, and the interpretation and Planned sensitivity analyses reporting of the results. To assess whether the results are robust to trial design The final results of the study will be presented to the and quality, sensitivity analyses will be performed on the collaborators for discussion. The main manuscript will primary outcome to test changes of the results after ex- be prepared by the STRIDER IPD Study Group. The re- clusion of trials with smaller sample size, centres with vised draft paper will be circulated for final comment smaller sample size, and trials/centres with high rate of and agreement prior to publication. Publications arising missing values. We will impute data for binary missing from these data will be authored by all members of the outcomes using various scenarios, namely best-best ana- STRIDER IPD Study Group on behalf of all other collab- lysis, worst-worst analysis, best-worst analysis, and worst- orators participating, who will be acknowledged within best analysis. the manuscript. Multiple comparisons Discussion A very large number of outcomes are being investigated Severe early-onset FGR due to placental insufficiency is in this study, which increases the risk of observing ‘false associated with a high risk of perinatal morbidity with positive’ results due to multiplicity. However, all out- long-lasting sequelae and mortality. Placental insuffi- comes are important in giving a full clinical picture that ciency is the result of abnormal formation and function considers the benefits and risks to both mothers and in- of the placenta (placentation) with inadequate remodel- fants. We do not plan formal statistical adjustment of P ling of the maternal spiral (uteroplacental) arteries. values to account for multiple comparisons due to the There is currently no therapy available with demon- non-independence of outcomes in this study. Results strated effectiveness, making monitoring and timely de- will be interpreted with caution. Decisions as to whether livery of the child the only treatment option. Both the intervention should be recommended will be based in vitro and in vivo evidence suggests sildenafil citrate as Ganzevoort et al. Systematic Reviews 2014, 3:23 Page 8 of 9 http://www.systematicreviewsjournal.com/content/3/1/23 a therapeutic strategy to improve uteroplacental blood collaborations, and reviewed the manuscript. All authors read and approved the final manuscript. flow, fetal growth, and meaningful outcomes in FGR pregnancies. Acknowledgements On the basis of preliminary research, some centres are New Zealand/Australia: funded by the Health Research Council (New already adopting the treatment with sildenafil. There is, Zealand). UK/Ireland study: funded by the Medical Research Council (United Kingdom). however, significant uncertainty regarding the true health benefits, if any. Moreover, potential harm is not yet ex- Author details cluded. Fetotoxicity has not been described yet, but is the- Department of Obstetrics and Gynecology, Academic Medical Centre, Room: H4-278, PO BOX 22660, 1100 DD Amsterdam, The Netherlands. oretically possible. For these reasons, evidence from Department of Women’s and Children’s Health, Liverpool Women’s Hospital, randomised clinical trials is needed before widespread im- University of Liverpool, Crown Street, Liverpool L8 7SS, UK. Department of plementation [31]. Obstetrics and Gynaecology, University of British Colombia, Rm V3-339, 950 West 28th Avenue, Vancouver, BC V5Z 4H4, Canada. The Irish Centre for Given the sample size issues, no single randomised Fetal and Neonatal Translational Research, University College Cork, Cork, clinical trial will ultimately be likely to give the answers Ireland. Department of Obstetrics and Fetal Medicine, St George’s Hospital, to the questions posed in this study. For this reason, Blackshaw Road, London SW17 0QT, UK. Department of Neonatology, Academic Medical Centre, PO BOX 22660, 1100 DD Amsterdam, The simultaneous smaller and medium-sized randomised Netherlands. Department 7812, Copenhagen Trial Unit, Centre for Clinical clinical trials have been planned and funded. The present Intervention Research, Rigshospitalet, Copenhagen University Hospital, pre-planned systematic review with IPD meta-analysis will Blegdamsvej 9, DK 2100 Copenhagen, Denmark. Department of Obstetrics and Gynaecology, Academic Medical Centre, Room H4-213, PO Box 22700, allow more definite answers, including effect modifica- 1105 DE Amsterdam, The Netherlands. Liggins Institute, University of tions in relevant subgroups. Furthermore, the totality of Auckland, Level 2, Building 505, 85 Park Rd, Grafton, Private Bag 92019, evidence from IPD trials and trials reporting only aggre- Auckland 1142, New Zealand. gate data can be assessed and evaluated with GRADE Received: 30 October 2013 Accepted: 27 February 2014 methodology [32]. All weighing of results should take into Published: 11 March 2014 consideration risks of bias and aim to minimise these: sys- tematic errors (bias) due to methodology and design; risks References 1. Lee MJ, Conner EL, Charafeddine L, Woods JR Jr, Del Priore G: A critical of academic bias if there are conflicting results between birth weight and other determinants of survival for infants with severe studies from this GONet group and other studies; risks of intrauterine growth restriction. Ann N Y Acad Sci 2001, 943:326–339. random errors (play of chance) [33]. According to our 2. Ananth CV, Vintzileos AM: Maternal-fetal conditions necessitating a medical intervention resulting in preterm birth. Am J Obstet Gynecol 2006, knowledge, this is the first time a protocol for a systematic 195(6):1557–1563. review has been finalised and published before the first 3. Petersen SG, Wong SF, Urs P, Gray PH, Gardener GJ: Early onset, severe larger trial has been launched on a topic. With the All- fetal growth restriction with absent or reversed end-diastolic flow velocity waveform in the umbilical artery: perinatal and long-term Trials initiative and other initiatives for transparency in outcomes. Aust N Z J Obstet Gynaecol 2009, 49(1):45–51. clinical research, we hope that the number of trials provid- 4. Batton DG, DeWitte DB, Espinosa R, Swails TL: The impact of fetal ing IPD for this systematic review will be large. compromise on outcome at the border of viability. Am J Obstet Gynecol 1998, 178(5):909–915. 5. Wood NS, Marlow N, Costeloe K, Gibson AT, Wilkinson AR: Neurologic and Abbreviations developmental disability after extremely preterm birth. EPICure Study FGR: Fetal growth restriction; GONet: Global obstetrics network; Group. N Engl J Med 2000, 343(6):378–384. IPD: Individual participant data; NICU: Neonatal intensive care unit; 6. Larroque B, Ancel PY, Marchand-Martin L, Cambonie G, Fresson J, Pierrat V, STRIDER: Sildenafil therapy in dismal prognosis early-onset intrauterine Rozé JC, Marpeau L, Thiriez G, Alberge C, Bréart G, Kaminski M, Marret S, growth restriction; UBC: University of British Columbia. Epipage Study group: Special care and school difficulties in 8-year-old very preterm children: the Epipage cohort study. PLoS One 2011, 6(7):e21361. Competing interests 7. Ganzevoort W, Rep A, De Vries JI, Bonsel GJ, Wolf H, PETRA-investigators: None of the authors have competing interests to report. Authors PB, ZA, LK, Prediction of maternal complications and adverse infant outcome at AP, WG, BM, AvW, and PvD are involved in successful/ongoing funding admission for temporizing management of early-onset severe hypertensive applications for the studies that form the basis of this IPD (Health Research disorders of pregnancy. Am J Obstet Gynecol 2006, 195(2):495–503. Council, New Zealand; UK Medical Research Council, United Kingdom; 8. Guellec I, Lapillonne A, Renolleau S, Charlaluk ML, Roze JC, Marret S, Vieux R, ZonMW, The Netherlands; and CIHR, Canada). Monique K, Ancel PY, EPIPAGE Study Group: Neurologic outcomes at school age in very preterm infants born with severe or mild growth Authors’ contributions restriction. Pediatrics 2011, 127(4):e883–e891. WG participated in the study design and drafted the manuscript. ZA 9. Walther FJ, Den Ouden AL, Verloove-Vanhorick SP: Looking back in time: participated in the design of the study and reviewed the manuscript. PvD outcome of a national cohort of very preterm infants born in The developed the hypotheses, participated in the design of the study, instigated Netherlands in 1983. Early Hum Dev 2000, 59(3):175–191. the international collaborations, and reviewed the manuscript. LK developed 10. Stanley JL, Andersson IJ, Poudel R, Rueda-Clausen CF, Sibley CP, Davidge ST, the hypotheses, participated in the design of the study, instigated the Baker PN: Sildenafil citrate rescues fetal growth in the catechol-O-methyl international collaborations, and reviewed the manuscript. AP participated in transferase knockout mouse model. Hypertension 2012, 59(5):1021–1028. the design of the study and reviewed the manuscript. AvW participated in 11. Wareing M, Myers JE, O’Hara M, Baker PN: Sildenafil citrate (Viagra) the design of the study particularly neonatal outcome analysis and reviewed enhances vasodilatation in fetal growth restriction. J Clin Endocrinol the manuscript. CG participated in the design of the study, particularly the Metab 2005, 90(5):2550–2555. statistics section and reviewed the manuscript. BM participated in the design 12. Herraiz S, Pellicer B, Serra V, Cauli O, Cortijo J, Felipo V, Pellicer A: Sildenafil of the study and reviewed the manuscript. PB developed the hypotheses, citrate improves perinatal outcome in fetuses from pre-eclamptic rats. participated in the design of the study, instigated the international BJOG 2012, 119(11):1394–1402. Ganzevoort et al. Systematic Reviews 2014, 3:23 Page 9 of 9 http://www.systematicreviewsjournal.com/content/3/1/23 13. Samangaya RA, Mires G, Shennan A, Skillern L, Howe D, McLeod A, Baker PN: A randomised, double-blinded, placebo-controlled study of the phosphodiesterase type 5 inhibitor sildenafil for the treatment of preeclampsia. Hypertens Pregnancy 2009, 28(4):369–382. 14. von Dadelszen P, Dwinnell S, Magee LA, Carleton BC, Gruslin A, Lee B, Lim KI, Liston RM, Miller SP, Rurak D, Sherlock RL, Skoll MA, Wareing MM, Baker PN, Research into Advanced Fetal Diagnosis and Therapy (RAFT) Group: Sildenafil citrate therapy for severe early-onset intrauterine growth restriction. BJOG 2011, 118(5):624–628. 15. GONet. Available from: http://www.globalobstetricsnetwork.org. 16. Broeze KA, Opmeer BC, van der Veen F, Bossuyt PM, Bhattacharya S, Mol BW: Individual patient data meta-analysis: a promising approach for evidence synthesis in reproductive medicine. Hum Reprod Update 2010, 16(6):561–567. 17. Wetterslev J, Thorlund K, Brok J, Gluud C: Trial sequential analysis may establish when firm evidence is reached in cumulative meta-analysis. J Clin Epidemiol 2008, 61(1):64–75. 18. Thorlund K, Devereaux PJ, Wetterslev J, Guyatt G, Ioannidis JP, Thabane L, Gluud LL, Als-Nielsen B, Gluud C: Can trial sequential monitoring boundaries reduce spurious inferences from meta-analyses? Int J Epidemiol 2009, 38 (1):276–286. 19. Thorlund K, Imberger G, Walsh M, Chu R, Gluud C, Wetterslev J, Guyatt G, Devereaux PJ, Thabane L: The number of patients and events required to limit the risk of overestimation of intervention effects in meta-analysis–a simulation study. PLoS One 2011, 6(10):e25491. 20. Wetterslev J, Thorlund K, Brok J, Gluud C: Estimating required information size by quantifying diversity in random-effects model meta-analyses. BMC Med Res Methodol 2009, 9:86. 21. User Manual for Trial Sequential Analysis (TSA) from the Copenhagen Trial Unit, Center for Clinical Intervention Research, Rigshospitalet, Copenhagen, Denmark. http://www.ctu.dk/tsa/. 22. WHO Trial Portal. http://apps.who.int/trialsearch/. 23. Current Controlled Trials. http://www.controlled-trials.com. 24. The Australian and New Zealand Trials Register. http://www.anzctr.org.au. 25. ISRCTN Register. http://www.isrctn.org. 26. The Cochrane Handbook for Systematic Reviews of Interventions. http://www. cochrane.org/training/cochrane-handbook. 27. Savovic J, Jones H, Altman D, Harris R, Juni P, Pildal J, Als-Nielsen B, Balk E, Gluud C, Gluud L, Ioannidis J, Schulz K, Beynon R, Welton N, Wood L, Moher D, Deeks J, Sterne J: Influence of reported study design characteristics on intervention effect estimates from randomised controlled trials: combined analysis of meta-epidemiological studies. Health Technol As- sess 2012, 16(35):1–82. 28. Savovic J, Jones HE, Altman DG, Harris RJ, Juni P, Pildal J, Als-Nielsen B, Balk EM, Gluud C, Gluud LL, Ioannidis JP, Schulz KF, Beynon R, Welton NJ, Wood L, Moher D,Deeks JJ,SterneJA: Influence of reported study design characteristics on intervention effect estimates from randomized, controlled trials. AnnInternMed 2012, 157(6):429–438. 29. Wood L, Egger M, Gluud LL, Schulz KF, Juni P, Altman DG, Gluud C, Martin RM, Wood AJ, Sterne JA: Empirical evidence of bias in treatment effect estimates in controlled trials with different interventions and outcomes: meta-epidemiological study. BMJ 2008, 336(7644):601–605. 30. Lundh A, Sismondo S, Lexchin J, Busuioc OA, Bero L: Industry sponsorship and research outcome. Cochrane Database Syst Rev 2012, 12, MR000033. 31. Jakobsen JCGC: The necessity of randomized clinical trials. Br J Med Med Res 2013, 3(4):16. 32. Grade Working Group. http://www.gradeworkinggroup.org. 33. Keus F, Wetterslev J, Gluud C, van Laarhoven CJ: Evidence at a glance: error matrix approach for overviewing available evidence. BMC Med Res Submit your next manuscript to BioMed Central Methodol 2010, 10:90. and take full advantage of: doi:10.1186/2046-4053-3-23 Cite this article as: Ganzevoort et al.: STRIDER: Sildenafil therapy in • Convenient online submission dismal prognosis early-onset intrauterine growth restriction – a protocol • Thorough peer review for a systematic review with individual participant data and aggregate • No space constraints or color figure charges data meta-analysis and trial sequential analysis. Systematic Reviews 2014 3:23. • Immediate publication on acceptance • Inclusion in PubMed, CAS, Scopus and Google Scholar • Research which is freely available for redistribution Submit your manuscript at www.biomedcentral.com/submit

Journal

Systematic Reviews – Springer Journals

Published: Mar 11, 2014

There are no references for this article.