Description of the condition
Thyroid dysfunction in pregnancy
Thyroid dysfunction in pregnancy has been associated with a range of adverse maternal and fetal/infant outcomes, including miscarriage, pre-eclampsia (high blood pressure and protein in the urine), preterm birth (birth before 37 weeks of gestation) and maternal thyroid dysfunction in the postnatal period (Stagnaro-Green 2011). Questions have also been raised about increased risks of cognitive dysfunction and other adverse neurodevelopmental outcomes for children born to mothers with thyroid dysfunction during pregnancy (Stagnaro-Green 2012). Thyroid dysfunction in pregnancy may be categorised as: hypothyroidism (subclinical and overt), isolated hypothyroxinaemia, thyroid autoimmunity, and hyperthyroidism (including Graves' disease and gestational thyrotoxicosis) (Negro 2011a). Debate continues surrounding the most appropriate strategies to identify and subsequently manage women with thyroid dysfunction before, during and after pregnancy, to prevent the potential adverse consequences.
Normal pregnancy is associated with important and complex changes in maternal thyroid physiology and hormone production, and maternal and fetal thyroid hormone profiles change as gestation progresses. During the first trimester, normal thyroid function results in an increase in thyroxine (T4) and triiodothyronine (T3) production, and a subsequent inhibition of thyroid-stimulating hormone (TSH), in part due to the production of high concentrations of human chorionic gonadotrophin (hCG) (a hormone produced in pregnancy), which stimulates the TSH receptor (Krajewski 2011).
The relationship between maternal and fetal thyroid hormone production is particularly important during the first half of pregnancy. Thyroid hormone is critical for the early development and maturation of the fetal brain, and thus the maternal transfer of thyroid hormone is essential, especially during the first trimester (Cooper 2012). The fetal thyroid gland does not begin to synthesise thyroid hormone until approximately 12 to 13 weeks' gestation and prior to this time any requirement for thyroid hormone is reliant on the mother (Casey 2006).
Thyroid dysfunction during pregnancy, its associated adverse outcomes, as well as management strategies, have been widely researched over the last few decades; and clinical practice guidelines have been produced to assist health practitioners in the provision of appropriate care (De Groot 2012; Stagnaro-Green 2011). Ongoing debate however, continues regarding the most effective way to identify women with thyroid dysfunction prior to and during pregnancy, and also women who may be at risk of developing thyroid dysfunction during pregnancy.
Overt hypothyroidism (OH) is characterised by an underactive thyroid gland, or symptomatic thyroid hormone deficiency, defined by low free thyroxine (T4) and high TSH concentrations. The prevalence of overt hypothyroidism has been estimated between 0.2% and 1.0% (Negro 2011a). Confirmation of hypothyroidism is achieved through measuring serum TSH.
Globally, the most common causal factor for maternal hypothyroidism is iodine deficiency, with maternal iodine deficiency impairing the synthesis of maternal and fetal thyroid hormones. Iodine is an important substance for the thyroid gland to produce and secrete thyroid hormones (Krajewski 2011). Due to increased thyroid hormone production, increased renal excretion of iodine in addition to the iodine requirements of the fetus, iodine requirements for women during pregnancy are higher than for non-pregnant women. The World Health Organization recommends an iodine intake of 250 micrograms per day to meet the increased demand during pregnancy. In high-income countries where iodine deficiency is less common, autoimmune thyroiditis (or Hashimoto's disease) is a more common cause of maternal hypothyroidism (Stagnaro-Green 2011).
Untreated maternal OH may be associated with significant complications for a mother and her baby, including hypertension, pre-eclampsia, miscarriage, placental abruption (a complication of pregnancy, where the placental lining separates from the uterus) and postpartum haemorrhage (loss of blood following birth). Fetal or infant complications associated with untreated maternal OH include preterm birth and low birthweight (Stagnaro-Green 2012). Deficiency of thyroid hormone during critical periods of development has the potential to damage the nervous system, and while specific effects are not certain, evidence has accumulated to suggest thyroid hormone deficiency might be one of the causes of cerebral palsy (Hong 2008). A number of observational studies have linked maternal thyroid hormone concentrations in pregnancy with childhood neurodevelopmental outcomes; for example, in two studies, children born to mothers with untreated hypothyroidism during pregnancy, were shown to have significantly lower intelligence quotient (IQ) scores, when compared with control children (Haddow 1999; Mitchell 2004).
OH may be treated with levothyroxine, a thyroxine replacement, with the goal of maintaining a euthyroid state (thyroid hormones within a normal range) throughout the duration of pregnancy. Treatment is considered critical to reduce the complications associated with untreated disease, and while treatment has been consistently supported in clinical practice guidelines (De Groot 2012; Stagnaro-Green 2011), the relevant Cochrane review ('Interventions for clinical and subclinical hypothyroidism pre-pregnancy and during pregnancy') recently highlighted the lack of randomised controlled trial evidence in this area (Reid 2013).
Subclinical hypothyroidism (SH) refers to biological evidence of thyroid hormone deficiency in women who experience none, or very few clinical symptoms. SH is defined as a normal free T4 concentration with high TSH concentration. SH is considered to be the most common thyroid disorder to occur in pregnancy and the prevalence has been estimated between 1.5% and 4% (Negro 2011a).
SH during pregnancy has been associated with an increased incidence of adverse outcomes similar to those associated with OH, however, in contrast to OH, SH represents a less severe degree of thyroid dysfunction (Cooper 2012). Studies have associated untreated maternal SH to early loss of pregnancy, increased rates of placental abruption and preterm birth (Le Beau 2006). Pregnancy loss, or miscarriage, is often noted as one of the most common complications linked to SH. Data from a prospective cohort study by Benhadi and colleagues reported an increased risk of miscarriage, fetal death and neonatal death with increasing concentrations of TSH during pregnancy, even in healthy women with no overt thyroid disorder (Benhadi 2009).
Treatment of SH with levothyroxine is not universally accepted, with recommendations differing between professional organisations (Stagnaro-Green 2012), and no clear evidence to support treatment from the relevant Cochrane review (Reid 2013).
Isolated hypothyroxinaemia (IH) during pregnancy is defined by the detection of low serum-free thyroxine (FT4) with TSH levels within the normal range. Data suggest wide variation in the incidence of IH due to global differences in maternal iodine intake and the reliability of diagnostic tools (Negro 2011a). Several studies have demonstrated that IH, largely caused by iodine deficiency (Morreale 2000), like hypothyroidism, can be associated with negative effects on fetal brain development and thus neurodevelopmental outcomes; increased risks of cognitive, language and motor dysfunction, including cerebral palsy have been reported (Hong 2008). For example, low maternal FT4 has been associated with an increased risk of impaired psychomotor development for children at 10-month follow-up (Pop 1999) and delayed mental and motor function at 12 and 24 months (Pop 2003); increased risks of impaired motor and intellectual development at 25 to 30 months (Li 2010); and expressive language delay and nonverbal cognitive delay at 18 to 30 month follow-up have also been reported (Ghassabian 2014).
Maternal thyroid autoimmunity refers to the detection of thyroid antibodies against thyroperoxidase (TPO-Ab) and/or thyroglobulin (Tg-Ab) in combination with normal thyroid function. The estimated incidence of thyroid autoimmunity in women of reproductive age has been reported to be between 8% and 14% (Vissenberg 2012). The precise mechanisms that underpin thyroid autoimmunity are not known, however it has been hypothesised that the presence of antibodies are part of an undefined autoimmune response, or that they may reflect a subtle reduction in optimum thyroid function (Negro 2011b). Thyroid autoimmunity has been associated with an increased risk of the development of maternal hypothyroidism during pregnancy (van den Boogaard 2011), and postnatal thyroid dysfunction (Stagnaro-Green 2011). The detection of thyroid antibodies in the early stages of pregnancy confers a 33% to 50% chance of the woman developing postpartum thyroiditis (Stagnaro-Green 2004). Women with thyroid antibodies during pregnancy who recover from postpartum thyroiditis experience a 70% recurrence rate in future pregnancies, and additionally, are at a significantly higher risk of developing permanent hypothyroidism (Samuels 2012).
Thyroid autoimmunity during pregnancy has been associated with poor fertility outcomes, recurrent miscarriage and preterm birth (Thangaratinam 2011; van den Boogaard 2011). Thyroid autoimmunity in combination with high TSH concentrations early in gestation has also been associated with a four-fold increased risk of gestational diabetes for the mother (Karakosta 2012). For women with insulin dependent diabetes mellitus, the presence of TPO-Ab prior to pregnancy has been suggested to confer an increased risk of poorer glucose control and an increased risk of developing hypothyroidism (Fernandez-Soto 1997). While the exact relationships between the conditions are unclear, both thyroid autoimmunity and postpartum thyroiditis have been associated with depression (Stagnaro-Green 2004). In a recent study, pregnant TPO-Ab positive women were shown to have higher depressive symptoms during pregnancy, and higher depression, anger and total mood disturbance postpartum, regardless of the development of postpartum thyroiditis (Groer 2013).
Similar to risks for the child associated with hypothyroidism and hypothyroxinaemia, children born to mothers with thyroid autoimmunity have been shown to be at risk of adverse neurodevelopmental outcomes. Elevated concentrations of TPO-Ab in women between 16 to 20 weeks of gestation have been shown to be a predictor of poor motor skills and intellectual development in early childhood (Li 2010). Maternal thyroid autoimmunity has also been associated with an increased risk of behavioural problems in childhood, particularly attention deficit and hyperactivity (Ghassabian 2012).
Current clinical guidelines outline no specific treatment for women with detected thyroid autoantibodies, however encourage monitoring of serum TSH during pregnancy due to the increased risk of hypothyroidism (Stagnaro-Green 2011). The relevant Cochrane review identified no significant difference for the outcome of pre-eclampsia when levothyroxine treatment was compared with no treatment for women who were TPO-Ab positive and euthyroid during pregnancy, however, noted a reduction in preterm birth and miscarriage with levothyroxine treatment (Reid 2013). The review emphasised that due to the small number of trials contributing data to the review, at a moderate risk of bias overall, that there was insufficient evidence to make clear conclusions and to guide clinical practice.
Hyperthyroidism is characterised by an overactive thyroid gland and the release of excessive amounts of thyroid hormones. Hyperthyroidism during pregnancy is rare with reported incidences ranging from 0.1% to 0.4%, and most women who experience the condition are diagnosed prior to conception and pregnancy (Chang 2013). Causes of hyperthyroidism can be distinguished as immune or non-immune in origin (Negro 2011a), with the autoimmune condition Graves' disease accounting for an estimated 85% of cases (Le Beau 2006). Graves' disease (in which autoantibodies stimulate the thyroid TSH receptor) is characterised by varied clinical presentation including heat intolerance, shortness of breath, the presence of a goitre, ophthalmopathy (swollen tissue behind the eyes), and rarely, thyroid storm (an acute, life-threatening, hyper-metabolic state induced by excessive release of thyroid hormones) and congestive heart failure. Uncontrolled hyperthyroidism as a result of Graves' disease has been associated with an increased risk of perinatal complications, including miscarriage, stillbirth, preterm birth, fetal growth restriction, and pre-eclampsia (Stagnaro-Green 2012). Anti-thyroid drugs (such as propylthiouracil and carbimazole) may be used to treat hyperthyroid women during pregnancy with the goal of achieving a euthyroid state (Stagnaro-Green 2011). A recent Cochrane review ('Interventions for hyperthyroidism pre-pregnancy and during pregnancy') however found no completed randomised controlled trials for inclusion addressing management in this area (Earl 2013).
The most common cause of non-immune hyperthyroidism is gestational thyrotoxicosis, also referred to as transient gestational hyperthyroidism, caused by elevated concentrations of hCG beyond normal range in early pregnancy. Diagnosis is confirmed by suppressed or very low serum TSH concentrations in the presence of high T4 (Negro 2011a). Gestational thyrotoxicosis is commonly associated with hyperemesis gravidarum (or severe morning sickness) and characterised by severe nausea, early-onset vomiting, and dehydration during the first trimester of pregnancy. Gestational thyrotoxicosis has not been associated with poor obstetric or infant outcomes, however women affected during pregnancy may require hospitalisation for supportive treatment to manage dehydration and other symptoms (Stagnaro-Green 2012). Anti-thyroid drugs have not been recommended, as serum T4 generally returns to normal concentrations between 14 and 18 weeks' gestation (Stagnaro-Green 2011).
Description of the intervention
Given the prevalence of thyroid disorders in pregnancy, the significance of the short- and long-term adverse outcomes for both the mother and her baby, and the availability of potential management options (dependent on the type of disease and clinical scenario), the early identification of women with thyroid dysfunction prior to, during, or after pregnancy is of great importance (Negro 2011a; Vissenberg 2012).
Screening is a strategy used to identify an unrecognised disease in individuals; this can include those with pre-symptomatic or unrecognised symptomatic disease. Screening interventions are designed to detect disease early, to enable earlier intervention and management with the ultimate aim of improving health outcomes (NSC 2013). Currently, maternal thyroid disorders are largely detected based on clinical presentation of symptoms and further assessment of thyroid hormone concentrations. Potential screening tests for thyroid dysfunction during pregnancy (such as testing for serum TSH and/or the presence of TPO-Ab) are readily accessible and the improved sensitivity of new generation assays allow for the detection of extremely low serum TSH concentrations, often seen in early pregnancy (Glinoer 2010).
Current clinical practice guidelines differ in their recommendations for screening, however, generally advocate for a 'case finding' rather than a 'universal screening' approach (Chang 2013). Universal screening involves screening all pregnant women, whereas case finding involves screening a smaller group of individuals of perceived increased risk. Case finding strategies can also differ, with some guidelines, such as those from the American College of Obstetrics and Gynaecology and from the United States Society for Maternal-Fetal Medicine, recommending thyroid testing only in high-risk pregnant women who are symptomatic, have a personal history of thyroid disorders, have type 1 diabetes or another autoimmune disorder (ACOG 2007; SMFM 2012). Other guidelines such as those of the American Association of Clinical Endocrinologists, American Thyroid Association, and Endocrine Society Task Force recommend a more 'aggressive' case finding approach (De Groot 2012; Garber 2012; Stagnaro-Green 2011) and provide detailed criteria for targeted thyroid disease case finding in newly pregnant women or women planning pregnancy (example criteria: over 30 years; family history of thyroid dysfunction; goitre; personal history of thyroid dysfunction; prior head/neck irradiation; prior thyroid surgery; symptoms or signs suggestive of dysfunction; thyroid antibodies, primarily TPO-Ab; type 1 diabetes or other autoimmune diseases; infertility; prior miscarriage or preterm birth; iodine deficient population; medications and iodinated contrast media; morbid obesity (Stagnaro-Green 2011)). Within individual guidelines, recommendations have also been found to differ. The Endocrine Society Task Force highlight in their guideline the lack of agreement reached regarding screening, noting that some members supported a universal screening approach to test for serum TSH abnormalities in pregnancy (prior to the ninth week of pregnancy, or at the initial visit), while other members supported an aggressive case finding approach (De Groot 2012).
How the intervention might work
It has been suggested that screening based on a 'case finding' approach may fail to detect a large proportion of hypothyroid and hyperthyroid women, whose thyroid dysfunction would thus remain uncontrolled throughout their pregnancies (Chang 2013); subclinical hypothyroidism and thyroid autoimmunity in particular, present with few or no obvious clinical symptoms (Vissenberg 2012). Thus, debate currently ensues surrounding the ideal strategy to identify thyroid dysfunction in pregnancy, particularly regarding the role of universal screening for all pregnant women, versus case finding.
A number of studies have assessed the effectiveness of current guidelines that emphasise targeted case finding of women with thyroid dysfunction before or during pregnancy, and their results have provided some support for the consideration of a universal screening approach. In a sample of 400 women, Horacek et al reported that over half (55%) of pregnant women with abnormalities detected in thyroid-related tests would have been overlooked if only high-risk case finding criteria were applied (Horacek 2010). Similarly, Vaidya et al revealed, in a sample of 1560 pregnant women, that specific targeting of only high-risk women would miss an estimated one-third of women with OH or SH (Vaidya 2007). In an assessment of thyroid testing and dysfunction rates at a medical centre in Boston, United States, Chang et al concluded that targeted testing for thyroid dysfunction would have missed approximately 80% of women with hypothyroidism during pregnancy (Chang 2011).
While universal screening may have the potential to identify a larger proportion of pregnant women with thyroid dysfunction and therefore facilitate earlier intervention and management, the implications of such an approach must be considered, including significant costs associated with treatment, follow-up and monitoring; and the possibility of misinterpretation of thyroid function tests, leading to over diagnosis/misdiagnosis, and the initiation of inappropriate management strategies (Chang 2013). In regards to the cost-effectiveness of a universal screening approach (compared with case finding), two evaluation studies considering SH (Thung 2009), and autoimmune thyroid disease (Dosiou 2012) in pregnancy both indicated that universal screening would in fact be a cost-effective tool, under a range of circumstances.
Why it is important to do this review
There is clearly documented evidence of morbidity and mortality associated with thyroid dysfunction (both hyper- and hypothyroidism) in pregnancy; thyroid dysfunction is associated with increased risks of a multitude of adverse outcomes for both the mother and her infant in the short and long term. While two Cochrane reviews have assessed interventions for the treatment of thyroid dysfunction pre-pregnancy and during pregnancy (Earl 2013; Reid 2013), no systematic review has assessed the effects of different methods of screening for thyroid dysfunction during pregnancy (including universal screening versus case finding), for the identification and subsequent management of thyroid disease.
Observational data from a number of studies have suggested that a large proportion of women with thyroid dysfunction in pregnancy are overlooked when targeted case finding methods are utilised (Chang 2011; Horacek 2010; Vaidya 2007). It is therefore important to conduct this review to assess the effectiveness of different methods of screening for thyroid dysfunction in pregnancy on detection, subsequent management, health outcomes and costs of care. It is plausible that universal screening (as compared with case finding) may facilitate improved diagnosis and early management of thyroid disease in pregnancy, and thus may assist in reducing associated maternal and infant complications.
To assess the effects of different screening methods (and subsequent management) for thyroid dysfunction pre-pregnancy, during pregnancy and in the immediate postpartum period on maternal and infant outcomes.
Criteria for considering studies for this review
Types of studies
Randomised controlled trials, quasi-randomised controlled trials and cluster-randomised trials will be included. Cross-over trials will be excluded. We will include studies published in abstract form only, along with those published in full-text form.
Types of participants
Women, either pre-pregnancy or during pregnancy (including both singleton and multiple pregnancies), or in the immediate postpartum period (six weeks after birth). We will exclude women with a pre-existing diagnosis of thyroid dysfunction.
Types of interventions
We will include trials where any screening method (e.g. tool, program, guideline or protocol) for detecting thyroid dysfunction (including hypothyroidism, hyperthyroidism, and/or thyroid autoimmunity) pre-pregnancy or during pregnancy, or in the immediate postpartum period (six weeks after birth) was compared with no screening.
We will also include trials where two or more methods of screening were compared (e.g. case finding versus universal screening).
Types of outcome measures
- Diagnosis of thyroid dysfunction (as determined by individual trialists)
- Hypothyroidism (overt; subclinical; isolated hypothyroxinaemia)
- Hyperthyroidism (Graves' diseases; gestational thyrotoxicosis)
- Thyroid autoimmunity
- Preterm birth (less than 37 weeks' gestation)
Infant as child
- Neurosensory disability (any of cerebral palsy, blindness, deafness, developmental delay/intellectual impairment, at latest time reported)
- Overall clinical improvement in symptoms of
- thyroid autoimmunity
- Pharmacological treatment required to manage thyroid dysfunction
- e.g. levothyroxine treatment for hypothyroidism
- e.g. propylthiouracil or carbimazole treatment for hyperthyroidism
- Pregnancy-induced hypertension
- Gestational diabetes
- Glucose tolerance measures (e.g. glycated haemoglobin (HbA1c) concentration)
- Congestive heart failure
- Thyroid storm
- Mode of birth (normal vaginal birth, operative vaginal birth, caesarean section)
- Induction of labour
- Preterm labour
- Placental abruption
- Postpartum haemorrhage
- Weight change (e.g. excessive weight gain/loss in pregnancy)
- Quality of life
- Adverse effects associated with the intervention
- Postpartum thyroid dysfunction
- Postnatal depression
- Maternal death
- Death (defined as all fetal and neonatal deaths)
- Fetal death
- Neonatal death
- Respiratory distress syndrome
- Low birthweight
- Small-for-gestational age (defined as birthweight less that 10th centile)
- Neonatal intensive care unit (or special care unit) admission
- Abnormal thyroid function (e.g. neonatal hyperthyroidism; neonatal goitre; cretinism – defined as congenital hypothyroidism resulting in impaired physical and mental development)
- Other congenital malformations
Infant as child
- Cerebral palsy
- Blindness: corrected visual acuity worse than 6/60 in the better eye
- Deafness: hearing loss requiring amplification or worse
- Developmental delay/intellectual impairment
- Maternal length of hospital stay
- Neonatal length of hospital stay
- Cost of screening
- Costs of maternal and neonatal/infant care
*Unless specified, for all outcomes above, we will include outcome data reported according to outcome definitions determined/specified by the trialists
Search methods for identification of studies
We will contact the Trials Search Co-ordinator to search the Cochrane Pregnancy and Childbirth Group’s Trials Register.
The Cochrane Pregnancy and Childbirth Group’s Trials Register is maintained by the Trials Search Co-ordinator and contains trials identified from:
- monthly searches of the Cochrane Central Register of Controlled Trials (CENTRAL);
- weekly searches of MEDLINE;
- weekly searches of Embase;
- handsearches of 30 journals and the proceedings of major conferences;
- weekly current awareness alerts for a further 44 journals plus monthly BioMed Central email alerts.
Details of the search strategies for CENTRAL, MEDLINE and Embase, the list of handsearched journals and conference proceedings, and the list of journals reviewed via the current awareness service can be found in the ‘Specialized Register’ section within the editorial information about the Cochrane Pregnancy and Childbirth Group.
Trials identified through the searching activities described above are each assigned to a review topic (or topics). The Trials Search Co-ordinator searches the register for each review using the topic list rather than keywords.
Searching other resources
We will search the reference lists of all retrieved studies.
We will not apply any date or language restrictions.
Data collection and analysis
Selection of studies
Two review authors will independently assess for inclusion all the potential studies we identify as a result of the search strategy. We will resolve any disagreement through discussion or, if required, we will consult a third review author.
We will create a Study flow diagram to map out the number of records identified, included and excluded.
Data extraction and management
We will design a form to extract data. For eligible studies, at least two review authors will extract the data using the agreed form. We will resolve discrepancies through discussion or, if required, we will consult a third review author. We will enter data into Review Manager software (RevMan 2014) and check for accuracy.
When information regarding any of the above is unclear, we will attempt to contact authors of the original reports to provide further details.
Assessment of risk of bias in included studies
Two review authors will independently assess risk of bias for each study using the criteria outlined in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). We will resolve any disagreement by discussion or by involving a third assessor.
(1) Random sequence generation (checking for possible selection bias)
We will describe for each included study the method used to generate the allocation sequence in sufficient detail to allow an assessment of whether it should produce comparable groups.
We will assess the method as:
- low risk of bias (any truly random process, e.g. random number table; computer random number generator);
- high risk of bias (any non-random process, e.g. odd or even date of birth; hospital or clinic record number);
- unclear risk of bias.
(2) Allocation concealment (checking for possible selection bias)
We will describe for each included study the method used to conceal allocation to interventions prior to assignment and will assess whether intervention allocation could have been foreseen in advance of, or during recruitment, or changed after assignment.
We will assess the methods as:
- low risk of bias (e.g. telephone or central randomisation; consecutively numbered sealed opaque envelopes);
- high risk of bias (open random allocation; unsealed or non-opaque envelopes, alternation; date of birth);
- unclear risk of bias.
(3.1) Blinding of participants and personnel (checking for possible performance bias)
We will describe for each included study the methods used, if any, to blind study participants and personnel from knowledge of which intervention a participant received. We will consider that studies are at low risk of bias if they were blinded, or if we judge that the lack of blinding would be unlikely to affect results. We will assess blinding separately for different outcomes or classes of outcomes.
We will assess the methods as:
- low, high or unclear risk of bias for participants;
- low, high or unclear risk of bias for personnel.
(3.2) Blinding of outcome assessment (checking for possible detection bias)
We will describe for each included study the methods used, if any, to blind outcome assessors from knowledge of which intervention a participant received. We will assess blinding separately for different outcomes or classes of outcomes.
We will assess methods used to blind outcome assessment as:
- low, high or unclear risk of bias.
(4) Incomplete outcome data (checking for possible attrition bias due to the amount, nature and handling of incomplete outcome data)
We will describe for each included study, and for each outcome or class of outcomes, the completeness of data including attrition and exclusions from the analysis. We will state whether attrition and exclusions were reported and the numbers included in the analysis at each stage (compared with the total randomised participants), reasons for attrition or exclusion where reported, and whether missing data were balanced across groups or were related to outcomes. Where sufficient information is reported, or can be supplied by the trial authors, we will re-include missing data in the analyses which we undertake.
We will assess methods as:
- low risk of bias (e.g. no missing outcome data; missing outcome data balanced across groups);
- high risk of bias (e.g. numbers or reasons for missing data imbalanced across groups; ‘as treated’ analysis done with substantial departure of intervention received from that assigned at randomisation);
- unclear risk of bias.
(5) Selective reporting (checking for reporting bias)
We will describe for each included study how we investigated the possibility of selective outcome reporting bias and what we found.
We will assess the methods as:
- low risk of bias (where it is clear that all of the study’s pre-specified outcomes and all expected outcomes of interest to the review have been reported);
- high risk of bias (where not all the study’s pre-specified outcomes have been reported; one or more reported primary outcomes were not pre-specified; outcomes of interest are reported incompletely and so cannot be used; study fails to include results of a key outcome that would have been expected to have been reported);
- unclear risk of bias.
(6) Other bias (checking for bias due to problems not covered by (1) to (5) above)
We will describe for each included study any important concerns we have about other possible sources of bias.
We will assess whether each study was free of other problems that could put it at risk of bias:
- low risk of other bias;
- high risk of other bias;
- unclear whether there is risk of other bias.
(7) Overall risk of bias
We will make explicit judgements about whether studies are at high risk of bias, according to the criteria given in the Handbook (Higgins 2011). With reference to (1) to (6) above, we will assess the likely magnitude and direction of the bias and whether we consider it is likely to impact on the findings. We will explore the impact of the level of bias through undertaking sensitivity analyses - see Sensitivity analysis.
Measures of treatment effect
For dichotomous data, we will present results as summary risk ratio with 95% confidence intervals.
For continuous data, we will use the mean difference if outcomes are measured in the same way between trials. We will use the standardised mean difference to combine trials that measure the same outcome, but use different methods.
Unit of analysis issues
We will include cluster-randomised trials in the analyses along with individually-randomised trials. If clustering has not been taken into account in the trial's analysis, we will adjust the trial's sample sizes using the methods described in the Handbook using an estimate of the intracluster correlation co-efficient (ICC) derived from the trial (if possible), from a similar trial or from a study of a similar population. If we use ICCs from other sources, we will report this and conduct sensitivity analyses to investigate the effect of variation in the ICC. If we identify both cluster-randomised trials and individually-randomised trials, we plan to synthesise the relevant information. We will consider it reasonable to combine the results from both if there is little heterogeneity between the study designs and the interaction between the effect of intervention and the choice of randomisation unit is considered to be unlikely.
We will also acknowledge heterogeneity in the randomisation unit and perform a sensitivity analysis to investigate the effects of the randomisation unit.
We consider cross-over trials as inappropriate for this review question.
As infants from multiple pregnancies are not independent, we plan to use cluster trial methods in the analyses, where the data allow, and where multiples make up a substantial proportion of the trial population, to account for non-independence of variables (Gates 2004).
If we include studies using one or more treatment groups (multi-arm studies) where appropriate, we will combine groups to create a single pair-wise comparison. We will use methods described in the Handbook (Higgins 2011) to ensure that we do not double count participants.
Dealing with missing data
For included studies, we will note levels of attrition. We will explore the impact of including studies with high levels of missing data in the overall assessment of treatment effect by using sensitivity analysis.
For all outcomes, we will carry out analyses, as far as possible, on an intention-to-treat basis, i.e. we will attempt to include all participants randomised to each group in the analyses, and all participants will be analysed in the group to which they were allocated, regardless of whether or not they received the allocated intervention. The denominator for each outcome in each trial will be the number randomised minus any participants whose outcomes are known to be missing.
If we include studies where women were recruited preconception, for outcomes relating to pregnancy, we plan to take a pragmatic approach and include in the denominators only those women known to have become pregnant.
Assessment of heterogeneity
We will assess statistical heterogeneity in each meta-analysis using the T², I² and Chi² statistics. We will regard heterogeneity as substantial if an I² is greater than 30% and either a T² is greater than zero, or there is a low P value (less than 0.10) in the Chi² test for heterogeneity.
Assessment of reporting biases
If there are 10 or more studies in the meta-analysis, we will investigate reporting biases (such as publication bias) using funnel plots. We will assess funnel plot asymmetry visually. If asymmetry is suggested by a visual assessment, we will perform exploratory analyses to investigate it.
We will carry out statistical analysis using the Review Manager software (RevMan 2014). We will use fixed-effect meta-analysis for combining data where it is reasonable to assume that studies are estimating the same underlying treatment effect: i.e. where trials are examining the same intervention, and the trials’ populations and methods are judged sufficiently similar. If there is clinical heterogeneity sufficient to expect that the underlying treatment effects differ between trials, or if substantial statistical heterogeneity is detected, we will use random-effects meta-analysis to produce an overall summary, if an average treatment effect across trials is considered clinically meaningful. The random-effects summary will be treated as the average of the range of possible treatment effects and we will discuss the clinical implications of treatment effects differing between trials. If the average treatment effect is not clinically meaningful, we will not combine trials.
If we use random-effects analyses, the results will be presented as the average treatment effect with 95% confidence intervals, and the estimates of T² and I².
We will consider separately trials screening for specific categories of thyroid dysfunction (i.e. hypothyroidism, hyperthyroidism) versus those trials screening for any thyroid dysfunction.
We will consider the following comparisons.
- Any screening method (tool, guidelines, protocol) versus no screening.
- One method of screening versus a different method of screening (such as case finding versus universal screening).
Subgroup analysis and investigation of heterogeneity
If we identify substantial heterogeneity, we will investigate it using subgroup analyses and sensitivity analyses. We will consider whether an overall summary is meaningful, and if it is, use random-effects analysis to produce it.
We plan to carry out the following subgroup analyses:
- gestational age at screening for dysfunction used in the trial (e.g. preconception versus first trimester versus second trimester versus third trimester versus immediate postpartum period);
- type of screening method/protocol used in the trial (e.g. case finding (based on symptoms/history only) versus 'aggressive' case finding (based on a range of risk factors));
- baseline risk for thyroid dysfunction of women in the trial (e.g. high risk for thyroid dysfunction (i.e. symptomatic women; women with a family history; other risk factors) versus low risk for dysfunction).
Primary outcomes will be used in subgroup analysis.
We will assess subgroup differences by interaction tests available within RevMan (RevMan 2014). We will report the results of subgroup analyses quoting the Chi² statistic and P value, and the interaction test I² value.
Sensitivity analyses will be conducted to explore the effects of trial quality and trial design on the outcomes. We will explore the effects of trial quality assessed by allocation concealment and random sequence generation (considering selection bias), by omitting studies rated as 'high risk of bias' (including quasi-randomised trials) or 'unclear risk of bias' for these components. We will investigate the effects of the randomisation unit (individual versus cluster) on the outcomes, and the impact of including studies with high levels of missing data. We will explore the effects of fixed-effect or random-effects analyses for outcomes with statistical heterogeneity, and the effects of any assumptions made such as the value of the ICC used for cluster-randomised trials. We will restrict this to the primary outcomes.
As part of the pre-publication editorial process, this protocol has been commented on by three peers (an editor and two referees who are external to the editorial team) and the Group's Statistical Adviser.
Contributions of authors
Laura Spencer drafted the first version of this protocol, with Emily Bain's guidance and feedback. Tanya Bubner and Philippa Middleton made comments on subsequent drafts and contributed to the final version.
Declarations of interest
Sources of support
- ARCH, The Robinson Research Institute, The University of Adelaide, Australia.
- National Health and Medical Research Council, Australia.