Blood manganese level and gestational diabetes mellitus: a systematic review and meta-analysis

Abstract Background Previous studies evaluating the relationship between blood manganese (Mn) level and gestational diabetes mellitus (GDM) in pregnant women showed inconsistent results. A systematic review and meta-analysis was therefore performed to investigate the above association. Methods Relevant observational studies were obtained by search of electronic databases including Medline, Embase, Cochrane Library and Web of Science from database inception to 10 March 2023. Two authors independently performed database search, literature identification and data extraction. A randomised-effects model was selected to pool the data by incorporating the influence of potential heterogeneity. Subgroup analysis was performed to evaluate the influence of study characteristics on the results of the meta-analysis. Results Six datasets from five observational studies, involving 91,249 pregnant women were included in the meta-analysis. Among the participants, 3597 (3.9%) were diagnosed as GDM. Overall, pooled results showed that a high blood level of Mn was associated with a higher risk of GDM (compared between women with highest versus lowest category blood Mn, odds ratio: 1.31, 95% confidence interval: 1.19–1.44, p < .001) with no significant heterogeneity (p for Cochrane Q-test = 0.93, I2 = 0%). Subgroup analyses according to study design, mean maternal age, matrix or methods for measuring blood Mn, and the incidence of GDM also showed consistent results (p for subgroup difference all >.05). Conclusions Results of the meta-analysis suggest that a high blood Mn level may be a risk factor of GDM in pregnant women. Studies are needed to determine the underlying mechanisms, and to investigate if the relationship between blood Mn level and GDM is dose-dependent. Plain Language Summary Changes of blood manganese (Mn) have been suggested to be involved in the pathogenesis of diabetes. However, the relationship between blood Mn level and gestational diabetes mellitus (GDM) in pregnant women remains undetermined. Our study represents the first systematic review and meta-analysis to investigate the potential association between blood Mn concentration and the risk of GDM. In this meta-analysis, we pooled the results of six datasets from five observational and showed that compared to pregnant women with the lowest category of blood Mn level, those with the highest category of blood Mn level were associated with a higher risk of GDM. These results suggest that a high blood concentration of Mn in pregnant women may be a risk factor of GDM.


Introduction
Gestational diabetes mellitus (GDM) is a common metabolic disorder in women with pregnancy (Choudhury andDevi Rajeswari 2021, Modzelewski et al. 2022).With the increased prevalence of obesity in the global population, millions of women are diagnosed with GDM during pregnancy annually all over the world (Jiang et al. 2022).
The growing body of evidence shows that GDM is associated with a variety of adverse maternal and neonatal outcomes (Saravanan 2020, Moon andJang 2022).Accumulating evidence demonstrates a correlation between GDM and a range of unfavourable outcomes for both mothers and neonates.A recent meta-analysis reveals that women with GDM who do not employ insulin therapy exhibit heightened likelihoods of undergoing caesarean sections, experiencing preterm deliveries, obtaining low one-minute Apgar scores, delivering macrosomic infants, and giving birth to infants with excessive size for their gestational age (Ye et al. 2022).Furthermore, despite the use of insulin, women with GDM still face increased odds of delivering infants with excessive size for their gestational age, neonatal jaundice or respiratory distress syndrome, or necessitating admission to a neonatal intensive care unit (Ye et al. 2022).Therefore, it is important for determining the risk factors and application appropriated strategies for the early prevention of GDM (Saravanan 2020).
Previous studies showed that changes in some blood trace elements concentrations may be related to the odds of GDM (McKeating et al. 2019).It has been shown that compared to those without GDM during pregnancy, women with GDM are associated with lower blood levels of selenium (Xu et al. 2022), magnesium (Ren et al. 2023), zinc (Fan et al. 2021), but higher blood levels of iron (Yang et al. 2022) and copper (Lian et al. 2021).Besides these trace elements, a change in blood manganese (Mn) has also been observed in patients with diabetes.Some studies showed a high blood Mn in patients with type 2 diabetes (T2D) as compared to those with normoglycaemia (Shan et al. 2016, Chen et al. 2022), while there are also studies that did not show the same results (Wang et al. 2016, Hansen et al. 2017).As for the association between blood Mn level and GDM, the evidence remains limited and has not been comprehensively evaluated (Lewicka et al. 2017).Therefore, a systematic review and meta-analysis were therefore performed in this study to investigate the correlation between blood Mn concentration and the risk of GDM in pregnant women.

Materials and methods
This meta-analysis was conducted and reported according to the Cochrane Handbook (Higgins et al. 2021) and the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA 2020) (Higgins et al. 2021, Page et al. 2021) guidelines.

Literature database search
In order to identify relevant studies, we searched Medline, Embase, Cochrane Library and Web of Science electronic databases according to the following terms: (1) ('manganese' OR 'Mn'); and (2) 'gestational diabetes mellitus' OR 'GDM' OR 'gestational diabetes' OR ('pregnant' OR 'pregnancy' OR 'gestational') AND ('hyperglycemia' OR 'diabetic' OR 'diabetes').The publications of the literatures were from database inception to 10 March 2023.We selected only English-language  studies involving human subjects.Additionally, we reviewed the reference lists of the original and review articles for potential inclusions that were not included in the original study.

Selection of eligible studies
Study inclusion was based on the PICOS criteria.The key exposure variable was a high blood Mn level during pregnancy and the outcome of interest of this study was the incidence of GDM.
1. P (participants): Pregnant women who visited the clinic for the measuring of blood Mn level; 2. I (intervention/exposure): Women with the highest category of blood Mn level during pregnancy; cutoffs used to differentiate high versus low levels of blood Mn were consistent with original study criteria; 3. C (control/comparator): Women with the lowest category of blood Mn level during pregnancy; 4. O (outcome): Oral glucose tolerance test (OGTT)-detected GDM during pregnancy based on original study criteria; reported risk estimates and the corresponding 95% confidence intervals (CIs) for the association between blood Mn and GDM, or these data could be accurately calculated.

S (study design): Research involving observational studies
with longitudinal follow-up, such as nested case-control studies, cohort studies and post hoc analyses of clinical trials; We excluded reviews, editorials, meta-analyses, preclinical studies and studies that did not measure blood Mn levels or did not report a GDM outcome.

Assessment of study quality and extraction of study data
Using Newcastle-Ottawa Scales (NOSs), the quality of studies was assessed in three domains: study group definition, between-group comparability and outcome validation (Wells et al. 2010).The NOS incorporated nine criteria, and each study received one point if it met a certain criterion.According to this scale, one star represents the lowest quality level of the study and nine stars represent the highest quality level.Study data were extracted from original articles independently by two authors with a self-designed form for data abstraction, and study quality was assessed based on the inclusion criteria described above by both authors.In the event of discrepancies between the two authors, they were resolved through discussion.We extracted the following data from each included study: (1) study information (authors, countries, publication year and study design); (2) numbers of women included, maternal age at evaluation, blood sampling time, matrix and methods for measuring blood Mn level, and cutoffs for categorising blood Mn level; (3) diagnostic criteria for GDM, and number of women who had GDM; and (4) variables included in the multivariate regression analysis for the association between blood Mn concentration and GDM in pregnant women.

Statistical methods
Odds ratios (ORs), together with their 95% CIs, were used as the statistical variable indicating the relationship between blood Mn level and the risk of GDM, which were compared between women with highest versus the lowest category of blood Mn level.To stabilise variance and normalise data, a logarithmical transformation was performed on ORs and standard errors (SEs) using 95% CIs and p values.(Higgins et al. 2021).An evaluation of heterogeneity was carried out using the Cochrane Q-test, along with an estimation of the I 2 statistic (Higgins and Thompson 2002).If I 2 is greater than 50%, heterogeneity is considered significant.Data synthesis was carried out using a randomised effects model, since this model accounts for potential heterogeneity between studies and provides a more generalised conclusion (Higgins et al. 2021).To evaluate the stability of the findings, we conducted sensitivity analyses by excluding one dataset at a time.Subgroup analysis was performed to evaluate the association between blood Mn concentration and GDM according to study design, mean maternal age, matrix and methods for measuring blood Mn, and the incidence of GDM in individual study.Based on the funnel plots, publication bias was assessed visually by examining their symmetry.Consequently, Egger's regression asymmetry test was performed in order to evaluate potential publication bias (Egger et al. 1997)

Literature database search and study identification
A brief summary of the initial literature search of the databases is shown in Figure 1.In total, 471 articles were identified.There were 342 studies remaining after excluding the duplicates.Through title and abstract screening, an additional 325 were excluded due to irrelevant content.We also reviewed full-texts of 17 remaining articles, but 12 were further excluded based on the reasons outlined in Figure 1.Finally, five observational studies (Soomro et al. 2019, Zhou et al. 2021, Li et al. 2022, Tatsuta et al. 2022, Wu et al. 2022) were subsequently studies in this meta-analysis.

Characteristics of the included studies
As shown in Table 1, three prospective cohort studies (Soomro et al. 2019, Li et al. 2022, Tatsuta et al. 2022), one retrospective study (Zhou et al. 2021) and one nested casecontrol study (Wu et al. 2022) were included in the meta-analysis.These studies were published between 2019 and 2023 and performed in France, China and Japan, respectively.Overall, 91,249 pregnant women were included in the metaanalysis.The mean maternal age at evaluation was 28-32 years.The timing of blood sampling was in the second trimester in three studies (Soomro et al. 2019, Zhou et al. 2021, Li et al. 2022), in the other two, it was in the first/second trimesters (Wu et al. 2022), or the second/third trimesters (Tatsuta et al. 2022).Whole blood Mn was measured in two studies (Soomro et al. 2019, Tatsuta et al. 2022), while plasma (Li et al. 2022, Wu et al. 2022) or serum Mn (Zhou et al. 2021) was measured in the other three studies.As for the methods for measuring blood Mn, the inductively coupled plasma mass spectrometry was used in three studies (Li et al. 2022, Tatsuta et al. 2022, Wu et al. 2022), the electrothermal atomic absorption spectrometry (Soomro et al. 2019) and polarography method (Zhou et al. 2021) were used in the other two studies.Different cutoffs were used to categorise the blood level of Mn in pregnant women, which involved the using of tertiles (Zhou et al. 2021), quartiles (Soomro et al. 2019, Li et al. 2022, Wu et al. 2022) and quintiles (Tatsuta et al. 2022) of Mn concentrations.For all the women included in these studies, a subsequent OGTT was performed for the diagnosis of GDM.In one of the included studies (Soomro et al. 2019), GDM was diagnosed according to the Carpenter and Coustan criteria (Kim et al. 2022), when there were �2 blood glucose concentrations greater than the following cut points: fasting ¼ 3.5 mmol/L, at 1 h ¼ 10.0 mmol/L, at 2 h ¼ 8.5 mmol/L and at 3 h ¼ 7.8 mmol/L.In the other studies, the International Association of the Diabetes and Pregnancy Study Groups Criteria were used (Metzger et al. 2010), if any of the following criteria were met: fasting blood glucose �5.1 mmol/L, at 1 h post-load blood glucose �10.0 mmol/L, and at 2 h post-load blood glucose �8.5 mmol/L.Accordingly, 597 (3.9%) women were diagnosed as GDM.Variables including maternal age, body mass index (BMI), smoking, socioeconomic factors and family history of diabetes, etc., were adjusted to a different degree among the included studies when the association between blood Mn level and GDM was estimated.The NOS of the included studies was 8-9 stars, suggesting good quality (Table 2).
No disagreement occurred between the two authors in data extraction and quality evaluation.

Meta-analysis results
Since one study reported outcomes according to BMI of the women (Li et al. 2022), these datasets were included into the meta-analysis independently.Overall, six datasets from five studies were included in the meta-analysis.Pooled results showed that a high blood level of Mn was associated with a higher risk of GDM (compared between women with highest versus lowest category blood Mn, OR: 1.31, 95% CI: 1.19-1.44,p < .001; Figure 2(A)) with no significant heterogeneity (p for Cochrane Q-test ¼ .93,I 2 ¼ 0%).In addition, sensitivity analyses by excluding one study at a time showed consistent results (p all <.001; Table 3).blood Mn (Figure 3

Publication bias
Figure S1 shows the funnel plots regarding the relationship blood Mn concentration and GDM.There was no evidence of Figure 3. Forest plots for the subgroup analysis of the association between blood Mn concentration and GDM in pregnant women.(A) Forest plots for the subgroup analysis according to the matrix for measuring blood Mn; (B) forest plots for the subgroup analysis according to the methods for measuring blood Mn; and (C) forest plots for the subgroup analysis according to the incidence of GDM of women in each study.
publication bias from visual inspection of the plots.Egger's regression test was not performed as there were only six datasets available to analyse.

Discussion
In this meta-analysis, we pooled the results of six datasets from five observational and showed that compared to pregnant women with the lowest category of blood Mn level, those with the highest category of blood Mn level were associated with a higher risk of GDM.Further sensitivity analysis by excluding one dataset at a time showed consistent results.Moreover, subgroup analysis according to study design, maternal age, matrix and measuring methods for blood Mn, and the incidence of GDM of women in each study also showed similar results.Taken together, results of the metaanalysis suggest that a high blood concentration of Mn in pregnant women may be a risk factor of GDM.
As far as we know, this is the first systematic review and meta-analysis that investigated the association between blood Mn concentration and the risk of GDM in pregnant women.The methodological strengths of the study may include the following aspects.First, an extensive literature search was performed in four commonly used electronic databases, aiming to retrieve the up-to-date literatures regarding the association between blood Mn level and risk of GDM in women with pregnancy.Second, all of the included studies were of longitudinal follow-up, which therefore could indicate a sequential relationship between increased blood Mn level and incidence of GDM.In addition, for all of the included studies, multivariate regression analysis was employed when the association between blood Mn concentration and the risk of GDM was evaluated.Accordingly, results of the meta-analysis suggested that the association between a high blood Mn and an increased risk of GDM may be independent of other risk factors for GDM, such as maternal age (Li et al. 2020), BMI (Santos et al. 2019) and family history of diabetes, etc. (Zhang et al. 2021).Finally, consistent results were obtained for multiple sensitivity and subgroup analyses, which further confirmed the stability and robustness of the finding.Overall, the above results indicate that a high blood level of Mn in pregnant women may be independently related to the higher risk of GDM.
Maternal obesity has been confirmed as an important risk factor for GDM (Santos et al. 2019).Indeed, among the included studies, only one study (Li et al. 2022) performed subgroup analysis according to the BMI of the patients, which showed consistent association between high blood Mn level and the risk of GDM.However, BMI was adjusted in the multivariate analyses among all of the included studies, which suggested that the association between blood Mn and the risk of GDM may be independent of maternal BMI.In addition, Mn is highly concentrated in the cellular elements of the blood, and compared to serum or plasma Mn, level of Mn in whole blood is higher (Schultze et al. 2014).However, subgroup analyses according to different matrix and methods for measuring Mn showed consistent results, which further confirmed the robustness of the association between a high blood level of Mn and the risk of GDM.
Currently, the mechanisms underlying the association between a high blood level of Mn and the risk of GDM remain to be elucidated.An early experimental study in streptozotocin-induced diabetic rats showed that Mn, as a cofactor for arginase, was elevated by four days post-streptozotocin injection, accumulated in the liver, involved in streptozotocin-induced diabetes (Bond et al. 1983).In addition, excessive Mn exposure and subsequently increased blood level of Mn could enhance the generation of reactive oxygen species (ROS) (Zhang et al. 2004).Because the expression of antioxidant defence genes is unusually low in pancreatic beta cells, which are extremely vulnerable to ROS-induced cellular injury (Gerber and Rutter 2017), it could be hypothesised that increment of systematic ROS may be a key process which mediates Mn related hyperglycaemia.Finally, excessive blood level of Mn may cause over-activated oxidative stress (Li and Yang 2018) and impaired mitochondrial dynamics (Smith et al. 2017), which have been both related to the pathogenesis of GDM (Joo et al. 2021, Kolac et al. 2021).Studies are needed to determine the molecular mechanisms underlying the association between Mn and GDM.
This study also has limitations.First, the number of studies available for the meta-analysis is small, and more prospective cohort studies are needed to validate the finding.In addition, the timings of blood sampling and cutoffs for categorising blood Mn level varied among the included studies, and the optimal cutoff value of blood Mn for the predicting the risk of GDM remains unknown.Studies are warranted in the future to resolve above issues.Besides, a dose-response relationship between blood Mn and GDM could not be determine on the basis of the current meta-analysis, because this study is based on studies with blood Mn analysed as categorised variables rather than continuous variables.Moreover, although multivariate regression analysis was used in all of the included studies, we could not exclude the possibility that there may be residual factors confounding the association.Finally, this meta-analysis was based on observational studies, and a causative relationship between high blood Mn and GDM could not be derived.

Conclusions
To sum up, results of the meta-analysis indicate pregnant women with high blood concentration of Mn may be associated with a higher risk of GDM.Studies are needed to determine the underlying mechanisms, and to investigate if the relationship between blood Mn level and GDM is dose-dependent.interpreted the data.Yingmei Sun drafted the manuscript.Yu Zhang revised the manuscript.Yingmei Sun and Yu Zhang approved the submission of the manuscript.

Figure 1 .
Figure 1.Flowchart of database search and study inclusion.

Figure 2 .
Figure2.Forest plots for the meta-analysis of the association between blood Mn concentration and GDM in pregnant women.(A) Forest plots for the overall metaanalysis; (B) forest plots for the subgroup analysis according to study design; and (C) forest plots for the subgroup analysis according to mean maternal age.
Finally, subgroup analyses according to study design (Figure 2(B)), mean maternal age (Figure 2(C)), matrix for measuring (A)), methods for measuring blood Mn (Figure 3(B)) and the incidence of GDM of each study (Figure 3(C)) also showed consistent results (p for subgroup difference all >.05).

Table 1 .
Characteristics of the included observational studies.

Table 2 .
Quality evaluation via the Newcastle-Ottawa Scale.

Table 3 .
Results of the sensitivity analysis.