Association between maternal folate status and gestational diabetes mellitus

Abstract Studies on the association between maternal folate status and gestational diabetes mellitus (GDM) have yielded inconsistent results. This meta‐analysis was performed to determine whether there may exist some association between maternal folate status and GDM. Unrestricted searches of PubMed, Web of Science, Cochrane, and Embase were conducted. All relevant studies on the association between maternal folat status and GDM risk were screened. The standardized mean difference (SMD) with 95% CIs was used to determine the association between maternal folate and GDM. Odds ratios (ORs) with 95% confidence intervals (CIs) were calculated using random‐effects models to assess the impact of maternal folate status on GDM risk. 12 studies were included. The overall data revealed that compared with the non‐GDM group, women with GDM had higher level of folate (SMD 0.41, 95% CI 0.07 to 0.21, I2 = 17.2%) in second or third trimester. We also found that maternal high folate status may be associated with increased risk of GDM (OR 2.16, 95% CI 1.70 to 2.74, I2 = 0.0%). Compared with non‐GDM group, women with GDM are prone to higher folate level. Moreover, high maternal folate status may predict a higher risk of GDM. As the number of included studies was limited, further large population studies are needed in the future.

, with a recommended daily dose of 400 micrograms (µg) from prepregnancy (4-12 weeks) until the end of the first trimester of pregnancy (8-12 weeks) in the majority (69.4%) of countries (Gomes et al., 2016). Surprisingly, many studies have found a link between maternal folate status and the development of GDM. However, studies have produced inconsistent conclusions regarding the association between maternal FA status and the risk of GDM (Cheng et al., 2019;Huang et al., 2019;Zhu et al., 2016). A prospective cohort (n = 3,747) demonstrated a positive association between daily FA supplementation in the first trimester and the risk of GDM (Zhu et al., 2016).
Consistent with this, high-dose FA supplementation in early pregnancy was related to an increased risk of GDM in a Chinese cohort (n = 4,353) . However, a large prospective cohort (n = 20,199) indicated that prepregnancy habitual intake of FA supplementation was negatively associated with GDM risk in the USA . Considering the inconsistent results mentioned above, determining the relationship between maternal FA status and GDM risk is urgent.
FA must be used during pregnancy to prevent NTDs in offspring, yet excessive use of FA may be associated with GDM risk (Cheng et al., 2019;Huang et al., 2019;Zhu et al., 2016). Together, these factors make it necessary to determine the role of maternal folate status in GDM development. This is the first meta-analysis to explore the association between maternal folate level and GDM.

| Article search strategy
This meta-analysis was rigorously projected according to the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) statement as described previously (Moher et al., 2009). The PROSPERO number is CRD42020191317. In this study, four electronic databases relevant to folate and GDM, including PubMed, Embase, Cochrane Library, and Web of Science, were searched. All published articles covering the relationship between folate and GDM were searched. The following search items were used: ("folic acid" OR "folate" OR "vitamin M" OR "vitamin B9" OR "B9, vitamin" OR "pteroylglutamic acid" OR "folvite" OR "folacin") and ("GDM" OR "gestational diabetes" OR "diabetes, pregnancyinduced" OR "diabetes, pregnancy induced" OR "pregnancy-induced diabetes" OR "diabetes mellitus, gestational"). Additional papers were identified by manual searching and citation tracking.

| Selection criteria
Two reviewers (YY and ZC) independently reviewed all the included studies, and the determination of eligible studies was finalized.
Disagreements were settled through consensus or the help of a third reviewer (JZ). All the articles to be included in this meta-analysis met the following criteria: (a) Studies with information on pregnancy folate status and GDM; (b) Studies with two groups, including a GDM and control group; and (c) Studies published in English. Articles were excluded if they met the following criteria: (a) Participants with multiple pregnancies or pre-existing diabetes (type 1 diabetes or type 2 diabetes); (b) Meeting abstracts that contained insufficient information for assessment; or (c) Studies that did not have available information or usable data for the purpose of this meta-analysis.

| Data extraction
All relevant articles were guided in EndNote X9 software and reviewed independently by two authors (YY and ZC). Discrepancies between authors were settled with the help of a third reviewer (JZ).
The following information was extracted by two independent investigators from the finalized study: author, year, country, study design, sample size, number of GDM, age (year), number of GDM cases, GDM criteria, and Newcastle-Ottawa Scale (NOS). All the extracted data were then imported into Excel software.

| Quality assessment of studies
The quality of the included studies was assessed by the NOS (Stang, 2010). We assessed the quality of all relevant studies in accordance with the type of study, sample size, participant selection, representativeness of the sample, adequacy of follow-up, comparability (exposed-unexposed or case-control), and method of ascertainment for cases and controls. A study with 6 or more scores was defined as high-quality.

| Definition of high folate status
So far, there is no accepted definition of high folate status. The World Health Organization (WHO) has recommended that RBC folate level greater than or equal to 906 nmol/L be the appropriate concentration for the prevention of neural tube defects (Cordero et al., 2015). Additionally, the first definition of serum high folate status was proposed by the WHO "based on the assay's upper limit capabilities without dilutions, and not on the biological implications for health" (WHO, 2015

| Statistical analysis
All statistical analyses were performed using Stata (Version 13.0).
The OR and 95% CI were used for discontinuous data. The heterogeneity among all studies was assessed by Q test and I2 statistics. All results were analyzed by a random-effects model. Sensitivity analysis was performed by excluding each study one by one to evaluate the credibility of pooled results. Begg's and Egger's tests and funnel plots were used to detect potential publication bias, with a pvalue < .05 suggesting the presence of bias. The trim-and-fill method was further conducted to obtain an adjusted effect size when encountering publication bias.
Among the 12 studies included in this analysis, there were

| Association between maternal folate status and GDM
As an indicator of maternal FA status, studies focusing on the concentration of maternal folate in GDM were analyzed. Nine studies were eligible for this part; among them, 14 results could be extracted to explore whether there was an association between maternal folate status and GDM (Table 2). Among the included nine studies, most of them performed GDM diagnosis during 24-28 weeks, with 3 exceptions. For comparison, all units of folate concentration were converted to nmol/L and all data presented as the medians and interquartile range were converted to the means and standard deviation according to the scenario 3 provided by Wan (Wan et al., 2014).
The overall data showed that compared with the normal glucose tolerance, the folate concentration in second or third trimester was slightly higher in women who developed GDM (random-effects pooled SMD [95% CI] 0.14 [0.07 to 0.21]; I 2 = 17.2%; Figure 2). Besides, two subgroup analyses were further conducted on the basis of measurement folate status (serum or RBC) and time for folate measurement (second or third trimester), while all the subgroup did not change the inverse association between folate status and GDM (Figure 2a-b). The results for Egger's test (p = .512), Begg's test (p = .511), and funnel plot showed no publication bias ( Figure S1a).
Moreover, sensitivity analyses also demonstrated that our results were stable ( Figure S1b).

| Association between maternal high folate status and GDM risk
Six studies were included for this part; among them, 9 results could be extracted to explore the effects of high maternal folate status on the risk of GDM (Table 3). Since most of the studies in this section refer to vitamin B12, the subgroup based on vitamin B12 was F I G U R E 1 Flow diagram of the study selection process analyzed in this part. The overall results showed that high maternal folate was associated with higher GDM risk (random-effects pooled OR [95% CI] 2.16 [1.70 to 2.74]; Figure 3). The prespecified subgroup based on vitamin B12 (B12) level further lessened the heterogeneity with no observed heterogeneity in each group (Figure 3). The OR was higher in the group with high FA concentrations combined with insufficient vitamin B12 (OR 2.40, 95% CI 1.60 to 3.61) and lower (OR 2.15, 95% CI 1.57 to 2.94) in the group with high FA concentrations regardless of vitamin B12 level (Figure 3). In addition, high FA combined with normal vitamin B12 had no association with GDM risk (OR 1.42, 95% CI 0.61 to 3.30) (Figure 3). Both Begg's test (p = .348), Egger's test (p = .192) and a symmetric distribution funnel plots ( Figure S2a) indicated no publication bias. Moreover, the results of sensitivity analysis demonstrated that our analysis was stable ( Figure S2b).

| D ISCUSS I ON
Women with GDM have an increased risk for prenatal morbidity and a considerably elevated risk for T2DM after pregnancy (McIntyre et al., 2019). FA supplementation during prepregnancy and pregnancy is widely used to prevent NTDs worldwide (Atta et al., 2016), contributing to an increased maternal folate status in women (Patel & Sobczyńska-Malefora, 2017). Recently, side effects caused by FA overuse or high folate status have raised concerns. To determine the association between high folate status and GDM, we first conducted this meta-analysis that was pooled from 12 studies.

| Association between maternal folate status and GDM
Compared with pregnant women with normal glucose tolerance, GDM women have higher levels of folate in the second or third trimester ( Figure 2). Our findings also revealed a positive association between high maternal folate concentration and GDM risk ( Figure 3). The results of our subgroup analysis of vitamin B12 levels demonstrated that the combination of a low vitamin B12 status and high folate concentration in pregnancy makes women prone to GDM, according to the results of our hierarchical comparison, and that the combination of high folate levels and a vitamin B12-normal status may have no association with GDM risk (Figure 3). These results indicated that low vitamin B12 levels may be one of the primary mechanisms driving the effect of high folate status on GDM.

| Underlying mechanisms of the effects of high folate status on GDM
Although the mechanism underlying the adverse effects of high folate status on the risk of GDM remains unclear, several factors have been proposed. First, unmetabolized folic acid (UMFA) in blood is associated with the dysfunction of NK cells (Bayer & Fraker, 2017;Troen et al., 2006), leading to the imbalance of immune function and participating in the process of diabetes (Bonamichi & Lee, 2017;Chiba et al., 2016).
Additionally, vitamin B12 along with folate is essential for onecarbon metabolism, including the synthesis of lipids, protein, and DNA (Finer et al., 2014). Vitamin B12 deficiency prevents DNA synthesis by inhibiting the production of tetrahydrofolate, while impaired mitochondrial DNA synthesis is associated with insulin resistance in obese subjects (Zheng et al., 2015).
Another possible mechanism is that low vitamin B12 during pregnancy can contribute to the increased levels of adiposederived circulating microRNAs, which may participate in the development of insulin resistance (Adaikalakoteswari et al., 2017).
Moreover, high folate and low vitamin B12 levels lead to elevated levels of homocysteine (Selhub et al., 2007), which have been observed among women with GDM compared with controls (Cho et al., 2005;Guven et al., 2006). Hyperhomocystinemia also plays a harmful role in pancreatic β-cell metabolism and insulin secretion (Engel et al., 2014).

| Implication for practice and research
This meta-analysis may have both important clinical and research implications. First, we demonstrated that maternal folate status was higher in GDM pregnant women than normal, and high folate status was associated with increased GDM risk especially in B12insufficient conditions, indicating that understanding the role of abnormally high folate levels in pregnant women is necessary. On the research side, our meta-analysis indicated that low B12 level may be the critical factor for the negative influence of high folate status.
More research is needed to identify the synergistic effect of high folate and low B12 status on GDM. The optimal maternal folate levels should also be the focus of future research.

| CON CLUS IONS
In summary, our meta-analysis shows that a high folate concentration may be related to an increased GDM risk, especially when combined with vitamin B12 deficiency, indicating that high FA concentrations combined with low vitamin B12 may be a novel marker for predicting GDM risk. However, more large and welldesigned studies are needed to prove this finding and explore the synergistic effect of high folate and low B12 status on GDM development. F I G U R E 2 Forest plot for overall and subgroup analysis of standardized mean difference in folate levels between GDM cases and controls. (A) Subgroup analysis based on time for folate measurement (second or third trimester). (B) Subgroup analysis based on measurement folate status (serum or RBC folate)

| S TUDIE S INVOLVING HUMAN SUBJEC TS
Although the study involves human subjects, it is a meta-analysis based on evaluating published research data. Therefore, no ethical issues are involved.

CO N FLI C T S O F I NTE R E S T
The authors declare no conflict of interest.

Jingjing Zhang
https://orcid.org/0000-0002-9155-8979 F I G U R E 3 Forest plot for subgroup analysis (based on vitamin B12 status) of association between high folate status and GDM risk