Effect of Hyperthyroidism Control During Pregnancy on Maternal and Fetal Outcome: A Systematic Review and Meta-Analysis

Context Although the overt hyperthyroidism treatment during pregnancy is mandatory, unfortunately, few studies have evaluated the impact of treatment on reducing maternal and fetal outcomes. Objective This study aimed to demonstrate whether treatment to control hyperthyroidism manifested during pregnancy can potentially reduce maternal-fetal effects compared with euthyroid pregnancies through a systematic review with meta-analysis. Data Source MEDLINE (PubMed), Embase, Cochrane Library Central, LILACS/BIREME until May 2021. Study Selection Studies that compared, during the gestational period, treated women with hyperthyroidism versus euthyroid women. The following outcomes of this comparison were: pre-eclampsia, abruptio placentae, fetal growth retardation, gestational diabetes, postpartum hemorrhage, low birth weight, stillbirth, spontaneous abortions, premature birth. Data Extraction Two independent reviewers extracted data and performed quality assessments. Dichotomous data were analyzed by calculating risk differences (DR) with fixed and random effect models according to the level of heterogeneity. Data Synthesis Seven cohort studies were included. The results of the meta-analysis indicated that there was a lower incidence of preeclampsia (p=0.01), low birth weight (p=0.03), spontaneous abortion (p<0.00001) and preterm birth (p=0.001) favouring the euthyroid pregnant group when compared to those who treated hyperthyroidism during pregnancy. However, no statistically significant differences were observed in the outcomes: abruptio placentae, fetal growth retardation, gestational diabetes mellitus, postpartum hemorrhage, and stillbirth. Conclusions Our findings demonstrated that treating overt hyperthyroidism in pregnancy is mandatory and appears to reduce some potential maternal-fetal complications, despite there still being a residual risk of negative outcomes.


INTRODUCTION
Hyperthyroidism, considering all causes, occurs in 0.2 to 1.3% of the population in countries where iodine intake is sufficient. Graves' disease (GD) is the most frequent cause and has a high prevalence in women who often appear in reproductive age (1,2).
Thyrotoxicosis can manifest in pregnancy in three forms: gestational thyrotoxicosis, subclinical hyperthyroidism, and overt hyperthyroidism. Gestational hyperthyroidism is a short form of thyrotoxicosis caused by hCG's excessive stimulation of the thyroid gland. It is usually limited to the first trimester of pregnancy. It affects 1-3% of all pregnancies, especially in women with hyperemesis gravidarum and multiple gestations (3). Subclinical hyperthyroidism is defined by TSH under the standard limit with average T4/T3 values, respecting the TSH and T4/T3 pregnancy reference values by trimester (1,3). Overt hyperthyroidism is rare, occurring in only 0.2% of pregnancies, and GD is its most common cause (3,4).
Although there is consensus on the need for treatment of overt hyperthyroidism during pregnancy, only a few studies evaluated the impact of this treatment on maternal and fetal endpoints (3,(5)(6)(7)(8)(9)(10)(11). Most data are conflicting regarding pregnancy outcomes such as pre-eclampsia, growth restriction, low birth weight, miscarriage, and premature birth. Hence, we aimed to understand better the maternal-fetal effects of treating overt hyperthyroidism with anti-thyroid drugs (ATD) (methimazole and propylthiouracil) during pregnancy, reducing adverse outcomes. The well-known teratogenic effects of ATD are beyond the scope of this review.

METHODS
This study was conducted following the PRISMA statement (12). The research protocol was registered in the International Prospective Register of Systematic Reviews (http://www.crd. york.ac.uk/PROSPERO) under CRD42021242704.

Study Selection
Two reviewers performed independent eligibility assessments to select the studies using predefined inclusion and exclusion criteria. Any divergence was resolved by consensus or consulting a third reviewer. The inclusion criteria were (I) pregnant women who have been diagnosed and treated with hyperthyroidism during pregnancy and for whom at least one pregnancy outcome has been assessed and (II) randomized controlled trials (RCTs) or non-randomized trials (NRS) or prospective or retrospective cohort studies with ATD treatment in one comparison arm regardless of the patients' number. The exclusion criteria were: (I) non-human studies, (II) letters, reviews, case reports, editorials, (III) studies without full text, and (IV) studies from which the necessary data could not be extracted from the pooled results.

Quality Assessment
Study quality was assessed using the Newcastle-Ottawa scale to assess the quality of non-randomized studies in meta-analyses, and certainty assessment was performed using GRADE (13,14). Disagreements were discussed between the investigators until a consensus was reached.

Data Extraction
One reviewer extracted all relevant information from acceptable studies, including design, sample size, population details, recruitment process, hyperthyroidism exposure, method of treatment, and outcomes. If data were reported in separate metrics, extracted outcome data were converted to a standard metric to estimate treatment effects.

Statistical Analysis
Statistical analyses were performed using the Review Manager software, version 5.4 (RevMan 5.4; Cochrane Collaboration, Oxford, UK). Dichotomous data were analyzed by computing risk differences (RD) with fixedand random-effect models employed according to the level of heterogeneity. Sensitivity analysis with funnel plot for ≥50% heterogeneity was not performed because, as a rule of thumb, tests for funnel plot asymmetry should be used only when at least ten studies are included in the meta-analysis. Also, the power of the tests is low when there are fewer studies.

Study Selection
After searching five databases and exploring reference lists, 1,225 potential studies were identified. The studies were uploaded to Endnote, where duplicates were excluded. After the exclusions, seven studies contained enough data to be included in a metaanalysis ( Figure 1).

Quality Assessment
All seven studies were considered as high quality by the Newcastle-Ottawa scale, as they scored between 7 and 8 ( Table 1). Due to the high score, all studies were included in the systematic review and meta-analysis.

Studies Characteristics
All included studies were based on a retrospective cohort conducted in India, Hungary, Thailand, Israel, Finland, and Denmark. The sample sizes ranged from 400 to 1,062,862, and the mean age of studies ranged from 25.5 to 30. The definition of hyperthyroidism during pregnancy varied among the included studies, but the studies clearly expressed the treatment of these women with antithyroid drugs ( Table 2).

Study Findings
No study included the assessment of all eligible outcomes. Data describing the presence of pre-eclampsia cases were available in six out of the seven eligible trials studies; data on gestational diabetes mellitus in five studies; fetal growth retardation, stillbirth, and premature birth in four studies; abruptio placentae data were available in three studies and spontaneous abortion, postpartum hemorrhage, and low birth weight in only two studies.

DISCUSSION
According to the 2017 ATA guidelines, poor control of thyrotoxicosis is associated with pregnancy loss, pregnancyinduced hypertension, prematurity, low birth weight, intrauterine growth restriction, stillbirth, thyroid storm, and maternal congestive heart failure (3). Unfortunately, there are little data on the effect of controlling thyrotoxicosis during pregnancy on maternal outcomes. Moreover, the risk for adverse maternal outcomes in women who had overt hyperthyroidism treated with ATD during pregnancy differs in various studies, which may be due to differences in inclusion criteria, sample size, and study design (5)(6)(7)(8)(9)(10)(11).
The present meta-analysis compared almost 6,000 pregnant women treated for hyperthyroidism with 1.3 million euthyroid pregnant women, demonstrating that the treatment for hyperthyroidism and restoration of the euthyroid state can supposedly reduce the incidence of five essential outcomes: placental abruption, delayed fetal growth, gestational diabetes, postpartum hemorrhage, stillbirth.
Pre-eclampsia is a pregnancy complication characterized by high blood pressure and signs of damage to another organ system, most often the liver and kidneys. Pre-eclampsia incidence range from 2 -7,5%; some risk factors are hypertension, obesity, diabetes mellitus, age, and race (15,16). Hyperthyroidism is a well-known risk factor for pre-eclampsia, especially poorly controlled (5,17). In addition, hyperthyroidism could aggravate a preexisting condition (e.g., hypertension) by predisposing to pre-eclampsia, or it can even trigger pre-eclampsia. Our data show that the development of preeclampsia was 4% lower in the pregnancies of euthyroid women.  Gestational diabetes mellitus is also an important outcome with several impacts on maternal health. Obesity is the major risk factor for diabetes mellitus (15,16,18). Also, hyperthyroidism is a well-known cause of increased insulin resistance and glucose levels. Both hormonal and immunologic conditions are related to this phenomenon (19). This meta-analysis suggests that overt hyperthyroidism treatment mitigates the deleterious effect of excessive thyroid hormone on glucose metabolism, preventing an increased risk of gestational diabetes mellitus (5,6,(8)(9)(10).
Spontaneous abortion is a tragic situation during pregnancy. Chromosomal abnormality is the single most common cause involved in approximately half of all cases of early spontaneous abortion and is also related to stillbirth (20). In addition, uncontrolled hyperthyroidism during pregnancy is a risk factor, although the molecular mechanism underlying this association is still not well understood (3). Although small but significant, our results showed that euthyroid pregnant women had 3% fewer events of spontaneous abortions.
One of the critical consequences of uncontrolled hyperthyroidism during pregnancy is low birth weight, which may be directly or indirectly related to maternal hyperthyroidism (21). Maternal hyperthyroidism may reduce fetal nutrition or act as a predisposing factor for other conditions that will cause these outcomes, similar to pregnancy-induced hypertension and maternal congestive heart failure (22). In women with GD, maternal TRAb passage to the placenta can induce low birth weight (3). Although overt hyperthyroidism was medicated with ATD, and presumably it was controlled; euthyroid pregnant women had 8% lower occurrence of low birth weight.
Placental abruption is the early separation of the placenta from the lining of the uterus before completing the second stage of labor (23). The primary cause is impairment of the vascular structures that support the placenta (24). Hyperthyroidism can predispose to hypertension, one of the most critical risk factors of placental abruption (3). Correcting hyperthyroidism during pregnancy attenuates this predisposition, resulting in no differences between hyperthyroid-treated women and euthyroid controls (5,8,10).
Postpartum hemorrhage is the most common cause of maternal mortality worldwide, and the hyperthyroid state contributes to coagulation disorders, acting directly on the gene transcription of coagulation proteins and altering the clot's structure (25,26). In addition, hyperthyroidism may cause a hyper-dynamic state and favor bleeding. Our data demonstrated that controlling hyperthyroidism equals the risk of postpartum hemorrhage compared to the control group (5, 7).
The World Health Organization defines premature birth as births before 37 completed weeks of gestation (27). Premature birth rates range from 3 to 14% in low-risk pregnancies. Maternal conditions, including hyperthyroidism, and other conditions such as pre-eclampsia, pre-gestational and gestational diabetes, and cervical incompetence may increase this incidence (28,29). This meta-analysis showed that even with the treatment of hyperthyroidism, the occurrence of preterm birth was 3% lower in euthyroid pregnant women (5,7,9,10).
This meta-analysis shows an excess risk of hyperthyroidtreated women in four endpoints (pre-eclampsia, low birth weight, spontaneous abortion, and premature birth) in comparison to euthyroid women. The differences were slight, ranging from 3 to 8%, however significant. The deleterious effect of hyperthyroidism affects the three trimesters of gestation, in the first-trimester spontaneous abortion and during the second and third-trimester pre-eclampsia, low birth weight, and premature birth. The mechanisms involved may be several and may be directly or indirectly related to hyperthyroidism, as discussed above. Also, other causes can be related to these negative outcomes, such as other autoimmune conditions associated      with GD (30). Also, GD TRAb could have essential participation in negative outcomes. TRAb passage through the placenta and the action in the thyroid fetus`gland also can determine the increase of these outcomes (31)(32)(33), especially in the third trimester. None of the studies evaluated in the meta-analysis mentioned the TRAb titers during treatment. The control of hyperthyroidism and the decrease in TRAb titers do not always present in the same period (34); thus, TRAb can cause fetus' hyperthyroidism even with maternal thyroid levels normal. Despite our attempts to eliminate potential biases, this systematic review has limitations. First, because some studies did not specify if hyperthyroidism was treated, we had to exclude many patients from the final analyses (35)(36)(37)(38)(39)(40)(41)(42)(43)(44)(45)(46). Second, it is important to emphasize that in carrying out this meta-analysis, we only used studies that showed the treatment of overt hyperthyroidism during pregnancy. Unfortunately, some studies are not specific about the degree of control of hyperthyroidism, we presumed that an euthyroid state was reached and maintained in all treated patients. Also, some studies may have started treatment after the first trimester. Third, the quality of evidence evaluated by the GRADE tool showed a very low certainty of the evidence for all outcomes. The main weak point in the quality of evidence was the type of study design, leading to a high level of imprecision. In addition, we were unable to examine data on subclinical hyperthyroidism and gestational thyrotoxicosis, limiting this meta-analysis to overt hypothyroidism.
No meta-analysis or systematic review has been published on this topic to our knowledge. In conclusion, treatment of overt hyperthyroidism in pregnancy is mandatory and appears to reduce some potential maternal-fetal complications. However, there is a residual risk of negative results even when overt hyperthyroidism is treated. This information will help doctors and patients manage pregnancy, especially those who needed to treat hyperthyroidism during this process.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/supplementary material. Further inquiries can be directed to the corresponding author.

AUTHOR CONTRIBUTIONS
JA: made considerable contributions to the design and postulation of the study, the definition of technical content, literature research, data analysis, statistical analysis, manuscript preparation, drafting, writing, critical review, and approval of the manuscript final version for publication. WB: were involved in the data analysis, statistical analysis, manuscript preparation, writing, drafting, critical review for important intellectual content. LW: Manuscript preparation, writing, drafting, critical review for important intellectual content, and approval of the manuscript final version for publication. DV: provided support for the entire process of developing and reviewing this systematic review and approval of the manuscript final version for publication. All authors contributed to the article and approved the submitted version.