Association of Polyphenols Consumption with Risk for Gestational Diabetes Mellitus and Preeclampsia: A Systematic Review and Meta-Analysis

Gestational Diabetes Mellitus (GDM) and preeclampsia (PE) affects 6–25% of pregnancies and are characterized by an imbalance in natural prooxidant/antioxidant mechanisms. Due to their antioxidant and anti-inflammatory properties, polyphenols consumption during the pregnancy might exert positive effects by preventing GDM and PE development. However, this association remains inconclusive. This systematic review and metanalysis is aimed to analyze the association between polyphenol-rich food consumption during pregnancy and the risk of GDM and PE. A systematic search in MEDLINE, EMBASE, and Web of Science (Clarivate Analytics, London, United Kingdom) for articles dated between 1 January 1980 and July 2022 was undertaken to identify randomized controlled trials and observational studies evaluating polyphenol-rich food consumption and the risk of GDM and PE. The Newcastle-Ottawa Scale was used to evaluate the quality of these included studies. Twelve studies were included, of which eight articles evaluated GDM and four studied PE. A total of 3785 women presented with GDM (2.33%). No association between polyphenol consumption and GDM was found (ES = 0.85, 95% CI 0.71–1.01). When total polyphenol intake was considered, a lower likelihood to develop GDM was noted (ES = 0.78, 95% CI 0.69–0.89). Furthermore, polyphenol consumption was not associated with PE development (ES = 0.90, 95% CI 0.57–1.41). In conclusion, for both outcomes, pooled analyses showed no association with polyphenol-rich food consumption during pregnancy. Therefore, association of polyphenol intake with a decreased risk of GDM and PE remains inconclusive.


Introduction
Gestational Diabetes Mellitus (GDM) and preeclampsia (PE) are common medical complications during pregnancy, both showing a complex etiology, which is characterized by a genetic charge and a marked influence of lifestyle [1,2]. Depending on diagnostic criteria, among 6% to 25% of pregnant women can be affected by GDM [3,4]. This high prevalence, in turn, represents a higher risk of pregnancy-induced hypertension and PE [5].
The pathophysiology of GDM involves a multisystem insulin resistance which has several causes, including a genetic predisposition, an unhealthy lifestyle, and the action of some hormones produced during pregnancy [1]. The insulin resistance leads to an increase in blood glucose levels, allowing an excessive transport of glucose into the fetus through the placenta [6]. Moreover, insulin resistance leads to an increase in free fatty acids

Methods
This systematic review and meta-analysis was based on an a priori established study protocol that followed the Cochrane Handbook guidelines for Systematic Reviews. The results are reported according to Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) guidelines. PRISMA checklist is presented in Supplementary Material (Supplementary Table S1).

Search Strategy and Information Sources
The search strategy was previously prepared by the authors, which was curated and optimized by the Karolinska Institute University Library professionals ( Figure 1). The electronic search included the revision of MEDLINE, EMBASE, and Web of Science (Clarivate Analytics) between 1 January 1980 and July 2021 and updated in September 2022, combining, between others, the following Medical Subject Headings (MeSH-terms) and keywords: 'polyphenols sub-classes or polyphenol-rich foods' and 'gestational diabetes mellitus' or 'preeclampsia'. This search strategy allowed the identification of relevant studies reporting original data for the evaluation of polyphenol consumption and the risk of GDM and PE (Supplementary Table S2). Further, all the included articles were manually reviewed to identify any unidentified publications. No restrictions on language, method of publication, or geographical area were applied. manually reviewed to identify any unidentified publications. No restrictions on language, method of publication, or geographical area were applied.

Eligibility Criteria and Study Selection
Firstly, all the studies identified by the systematic search were compiled in the reference management software ENDNOTE X20. The authors RF, ST, and GJ filtered and selected the articles, and GC solved discrepancies when needed. The articles included in the meta-analysis were based on the following eligibility criteria: original studies from 1980 without language restrictions that assessed the polyphenol intake in pregnant women between the ages of 18-40 years and the risk of GDM or PE; (ii) studies of polyphenol-rich foods consumption by food frequency questionnaires or direct dietary intervention of polyphenols, curcumin, resveratrol, flavonoids,

Eligibility Criteria and Study Selection
Firstly, all the studies identified by the systematic search were compiled in the reference management software ENDNOTE X20. The authors RF, ST, and GJ filtered and selected the articles, and GC solved discrepancies when needed. The articles included in the meta-analysis were based on the following eligibility criteria: (i) original studies from 1980 without language restrictions that assessed the polyphenol intake in pregnant women between the ages of 18-40 years and the risk of GDM or PE; (ii) studies of polyphenol-rich foods consumption by food frequency questionnaires or direct dietary intervention of polyphenols, curcumin, resveratrol, flavonoids, quercetin, tannins, catechins, phenolic acid, hydroxybenzoic acid, hydroxycinnamic acid, anthocyanins, or polyphenol-rich foods (tea, berries, chocolate, coffee); (iii) randomized controlled trials (RCTs), observational studies (prospective cohort studies); (iv) studies providing odds ratios (ORs), relative risks (RRs), or hazard ratios (HRs) along with 95% confidence intervals (CIs) or sufficient data to calculate the effect size (ES). When the article provided more than one estimator, the most adjusted one was selected.
The exclusion criteria were: (i) studies with no control of polyphenol-rich foods intake; (ii) case reports, abstracts, and reviews of animal or in vitro studies; (iii) narrative reviews, meta-analysis, systematic reviews, book chapters and conference abstracts.

Quality Assessment
The quality assessment of the included studies (cohort and case-control) was assessed by authors RF and ST using the Newcastle-Ottawa Scale (NOS) for observational studies and case-control studies. A NOS score ≥7 was considered of high quality (Supplementary Table S3). The disagreements were resolved via mutual consideration with author GJ. No studies were excluded due to low scores. A Meta-analysis of Observational Studies in Epidemiology (MOOSE) was incorporated in supplementary material (Supplementary  Material Table S4).

Data Extraction
After checking and excluding studies with same population assessed, the data extraction included: (i) characteristics of the population (age and sample size, comparison group); (ii) data on trials (design, country of origin, duration of follow up, and risk of bias); (iii) exposure (polyphenols, polyphenol-rich foods evaluated in direct interventions or food frequency questionnaires); (iv) outcome measured in the studies (confounders assessed and the crude and adjusted estimates); (v) essential data including available raw data of incidence and proportions, cumulative incidence of the events of interest and HR with 95% CI.

Data Synthesis and Statistical Analysis
The pooled effect size (ES) was calculated by DerSimonian and Laird random effects meta-analysis and the Cochrane Q and I-squared statistics were used to assess the heterogeneity between the studies, I 2 > 75% being indicative of high heterogeneity, 50-75% of moderate, and <50% of low heterogeneity. In case of Q test, p < 0.10 was considered indicative of heterogeneity. Supplementary analyses were conducted based on total polyphenol intake or their subclasses (i.e., anthocyanins, isoflavones, anthocyanins enriched fruits, and fruits). All statistical analyses were performed with Stata MP 17.0 (Stata Corp., College Station, TX, USA).

Results
The systematic search on polyphenols consumption and pregnancy complications included 4202 articles with publication dates until July 2021, uploaded in September 2022. After filtering based on titles and abstracts, 39 studies were included for a full assessment ( Figure 1). Only eight studies were deemed eligible for risk assessment of GDM and four for PE (Tables 1 and 2, respectively). Supplementary Materials Table S5 shows excluded studies. All selected studies were written in English and were published between 2007 and 2021. No new articles were found in the search conducted in September 2022.    # When the upper or lower limit of the upper or lower interval was not specified (i.e., >40), the mean ± SD of age and BMI was calculated by using the lower or upper limit written in the article (i.e., if the highest interval in BMI was >40, the lower and the upper limit of that specific interval was 40). Abbreviations: GDM = gestational diabetes mellitus; PE = preeclampsia; OGTT = oral glucose tolerance test; Q = quartile; BMI = body mass index; d = day; hrs = hours; FFQ = food frequency questionnaire; HIV = human immunodeficiency virus.

Selected Studies
In studies assessing the association between polyphenol consumption and the risk of GDM, women were recruited from hospitals [24,26,27,30,31] or national cohort studies [25,28,29]. In articles assessing the risk of PE, three studies were hospital or care practices based [20][21][22] and one was a national cohort study [23]. Age and body mass index (BMI) of the participants in each article are shown by exposure, or by group of women affected (Tables 1 and 2).
The evaluation of the association of polyphenol-rich food consumption during pregnancy and PE risk was based in only four studies that met the eligibility criteria [20][21][22][23]. Triche and co-authors [20] compared low consumption of polyphenol-rich foods vs. high consumption, whereas the other three articles compared no consumption with high consumption to estimate PE risk [21][22][23]. Kawanishi and co-authors [23] determined the polyphenol rich-food intake (tea or coffee) by using a questionnaire during the second or third trimester of pregnancy after recognizing conception, while other authors evaluated the risk to develop PE, estimating the intake of polyphenol-rich foods since women became pregnant until delivery [20][21][22]. In both cases, the analysis considered a total of 90,091 individuals. One study was retrospective [21] and three were cohort studies [20,22,23]. As the article from Kawanishi et al. (2021) [23] showed two estimators, tea and coffee consumption and the risk of PE, two meta-analyses were performed. An in-person interview with a questionnaire designed for the study was used for three studies [20][21][22], while just one used self-administrated questionnaires [23]. The food frequency questionnaire was used for just one study to determine polyphenol food consumption [23], while the others used self-designed questionnaires. These studies were conducted in North America [20][21][22] and Asia [23]. The information of the exposure corresponded to chocolate [20,22], or coffee or tea consumption [21,23]. The quality assessment of included studies is presented in Supplementary Material Table S2.  (Figure 2A and Figure 2B, respectively). The supplementary analysis showed that only quantified total polyphenol intake or their subclasses had significant results ES = 0.78 (0.69-0.89), I 2 : 0%, p < 0.37, n = 3] (Supplementary Figure S1A-C). Further, when anthocyanins-enriched fruits or polyphenols from fruits were analyzed, no association with GDM was found (Supplementary Figure S1D,E).

Preeclampsia Risk Assessment
In the selected observational studies, the incidence of PE was 2.  Figure 3B). A sub-analysis excluding Saftlas et al. was performed because they included gestational hypertension between study participants [22]. In this case, no association between polyphenol consumption and the risk of PE was found [ES = 1.12 (0.97-1.29), I 2 : 0%, p = 0.532, N = 3] (Supplementary Figure S2).

Our Findings
No association was found between overall polyphenol consumption and the risk of GDM or PE in this meta-analysis of twelve studies enrolling 252,534 pregnant women. However, in subtypes analyses, a protective effect (RR = 0.78) was noted against GDM. Coffee or chocolate consumption was associated with a reduced risk of PE in two studies involving 88,073 pregnant women [22,23]. This might imply a possible protective effect of these nutrients that warrants further study. Gestational diabetes mellitus and pre-eclampsia are associated with increased risks of maternal and fetal adverse outcomes [2,[32][33][34]. It is still unclear whether a specific diet might decrease the risk of developing GDM and PE. In a prior systematic review and meta-analysis, the risk of GDM was 0.57 (0.41-0.79) in those studies evaluating Mediterranean diet but not with polyphenol-rich foods intake [35]. However, the positive association might have been taken with caution. It because the mentioned study included cross-sectional reports, a majority of studies were based on self-questionnaires prone to recall bias, and the studies evaluated exposure to Mediterranean diet (and not proper polyphenols intake) [35]. Another systematic review involving 18 studies found that low-carbohydrate, ethnicity-based, and low glycemic index diet might be associated with better maternal glycemic control parameters [36]. No prior meta-analyses examining the risk of PE with polyphenols intake was found.
Both GDM and PE are characterized by an imbalance of oxidative stress in pregnant women [7,8]. It is necessary to improve the strategies to prevent these diseases. Adopting healthier lifestyle changes like healthy diets and increase physical activity has been associated with blood glucose control. However, whether polyphenols in particular have more benefit associated with lower risk is unknown/unclear. The potential role of polyphenol-rich foods against GDM and PE is described in Figure 4. [35]. However, the positive association might have been taken with caution. It because the mentioned study included cross-sectional reports, a majority of studies were based on self-questionnaires prone to recall bias, and the studies evaluated exposure to Mediterranean diet (and not proper polyphenols intake) [35]. Another systematic review involving 18 studies found that low-carbohydrate, ethnicity-based, and low glycemic index diet might be associated with better maternal glycemic control parameters [36]. No prior metaanalyses examining the risk of PE with polyphenols intake was found.
Both GDM and PE are characterized by an imbalance of oxidative stress in pregnant women [7,8]. It is necessary to improve the strategies to prevent these diseases. Adopting healthier lifestyle changes like healthy diets and increase physical activity has been associated with blood glucose control. However, whether polyphenols in particular have more benefit associated with lower risk is unknown/unclear. The potential role of polyphenolrich foods against GDM and PE is described in Figure 4.

Why Is It Expected That Polyphenol-Rich Food Intake Protects against GDM?
During the last decades, a growing number of in vitro, preclinical and clinical studies showed a positive effect of polyphenols from fruits and vegetables on glucose homeostasis. The suggested mechanisms were as follows: inhibiting digestive enzymes; reducing

Why Is It Expected That Polyphenol-Rich Food Intake Protects against GDM?
During the last decades, a growing number of in vitro, preclinical and clinical studies showed a positive effect of polyphenols from fruits and vegetables on glucose homeostasis. The suggested mechanisms were as follows: inhibiting digestive enzymes; reducing activation of glucose transporter in small intestinal epithelial cells; decreasing the uptake of glucose into blood for limiting basolateral cell transport [37][38][39]; protecting pancreatic βcells function; increasing insulin sensitivity [40,41]. All the mentioned mechanisms should lead to decreased basal and post-intake plasma glucose levels, leading to improved glucose tolerance. Evidence for improving glucose management has been observed for polyphenol rich fruits [42] and for resveratrol [43,44] in type 2 diabetic patients. Improving glucose management in pregnant women with GDM is fundamental to reduce their risk for developing pregnancy complications. For example, the recent metanalysis by Li X et al. [45] shows that exercise intervention improves glucose management and reduces adverse pregnancy outcomes such as premature birth and macrosomia. Similar improvements were observed under the treatment with insulin or metformin [46]. Selected fruits enriched in polyphenols, such as berries, have a great potential to improve glycemic management [42,47] and this improvement in glucose management should lead to improvements of pregnancy and neonatal outcomes.
Improvements in lipid metabolism of pregnant mothers with GDM is expected to reduce the impact of GDM on offspring development, since maternal triglycerides and total cholesterol are independent risk factors for pregnancy complications, although other lipid metabolites should also be considered [52].
Another factor implicated in GDM is gut microbiota. Indeed, the abundance of some genera of bacteria is associated with glucose metabolism in GDM women [53] and gut microbiota could be mediating the increase in BMI of infants born to GDM women [54].
Interestingly, it is suggested that treatment with probiotics could improve insulin sensitivity, glycemic control, and pregnancy outcomes in GDM women [55][56][57]. In this line, treatment with polyphenols improves gut microbiota dysbiosis in both animal models and human studies, which is also related to the reduction of systemic inflammation that is common in GDM and PE [58,59]. Therefore, polyphenol-rich food consumption may improve gut dysbiosis in GDM women, thus improving glycemic control and pregnancy outcomes.

Why Is It Expected That Polyphenol-Rich Food Intake Protects against PE?
Polyphenols can exert benefits on blood pressure regulation by inducing vasodilation throughout two different pathways, which include increment of nitric oxide (NO) and vascular smooth muscular cells (VSMC) relaxation [60]. The antioxidant properties induced by polyphenols, especially flavonoids, are responsible for reducing reactive oxygen species being related to increased eNOS activity, which raise NO activity and content, whereas VSMC relaxation is mediated by inhibition of vascular Ca 2+ channels and modulation of vascular BK Ca , endothelial IK Ca, and SK Ca channels [60]. Despite the wide evidence of beneficial effects of polyphenols on vascular function, it is unknown whether they might have a role in preventing preeclampsia.

Why Is It Important to Find New Strategies to Manage GDM and PE?
GDM increases the risk of metabolic disturbances for the newborn and perinatal outcomes, including congenital abnormalities, macrosomia or large for gestational age, stillbirth, prematurity, jaundice, hypoglycemia and/or hyperinsulinemia at birth, respiratory distress syndrome, shoulder dystocia, and higher risk of obesity, type 2 diabetes, and impaired glucose intolerance during adulthood [32][33][34].
The treatment of GDM is fundamental to prevent pregnancy complications and perinatal morbidity [61]. Moreover, we should go further and consider that improving glycemic control will protect the mother and their offspring from the long-term consequences triggered by GDM [62]. Currently, the treatments for GDM are lifestyle modifications and the use of insulin or metformin. Metformin and insulin treatments during GDM appear to be similar regarding the consequences in the mother [63] and their offspring [64], but they are the only two treatments for GDM.
PE is one of the most prevalent pregnancy-associated diseases which impacts both pregnancy outcomes and the newborn's outcomes. Unfortunately, there is no other cure for preeclampsia than delivery. However, PE management is essential to decrease the impact of low nutrient and oxygen supply to the fetus and also prevent serious consequences of high blood pressure in the mother. In this sense, the use of resveratrol as a supplement for the treatment with nifedipine resulted in improving blood pressure control in patients with PE [65]. Moreover, plasma from patients consuming resveratrol or polyphenol-rich red grape juice decreases antioxidant markers and improves nitric oxide production in human umbilical vein endothelial cells (HUVEC) [66].

Strengths and Limitations
This is the first systematic review and metanalysis that evaluated the association of polyphenol-rich food subtype intake and the risk of PE. Moreover, it is the second metanalysis that evaluated the association of polyphenol intake and the risk of GDM. However, as discussed above, in the present metanalysis, we excluded articles concerned with the Mediterranean diet, which has other bioactive compound with healthy proprieties (i.e., n-3 unsaturated fatty acids, carotenoids and others). The advantage of our article is that we selected mainly prospective studies in which the diagnosis criteria were equivalent among articles. Here, we used studies in which the consumption of polyphenol rich foods was either estimated with food questionnaires or calculated, in which all included articles describe the follow-up and the exposure, in which the population was characterized, and in which control groups were used. In addition, we performed sub-analyses depending on polyphenol source and sub-classes of polyphenol.
Despite these, the metanalysis has several limitations including the heterogenicity of clinical studies regarding the source of polyphenol intake, the variability of questionnaires employed, and their accuracy in establishing a dosing of polyphenols consumed per day in the studied population. The lack of effects observed in both GDM and PE could be a result of the small number of studies and the in between studies' heterogenicity. In addition, we could not ascertain the intake of polyphenols previous to pregnancy, which could add variability to the results of each study. Even if the clinical effects of polyphenols could be short term, it is expected that the benefits occur under a chronic consumption.

Conclusions
Overall consumption of a polyphenol-rich diet is not associated with decreased risk for GDM or PE. A beneficial effect might be present with coffee and chocolate consumption. The need for more prospective and/or controlled studies finding bioactive compounds that might prevent GM and PE is nevertheless warranted.
Supplementary Materials: The following supporting information can be downloaded at: https:// www.mdpi.com/article/10.3390/antiox11112294/s1. Table S1: PRISMA checklist; Table S2: The electronic search string in different databases; Table S3: The Newcastle-Ottawa Scale (NOS) for assessing the quality of non-randomized studies in meta-analysis; Table S4: MOOSE checklist;  Table S5: Excluded, full text assessed studies with a reason for exclusion from the systematic review and meta-analysis; Figure S1: Total polyphenol intake or their subclasses and GDM association; Figure S2: Total polyphenol intake and PE association without gestational hypertension.