Effects of Anthocyanin-rich Berries on the Risk of Metabolic Syndrome: A Systematic Review and Meta-analysis

OBJECTIVE Metabolic syndrome (MetS) can lead to fatal complications, including cardiovascular disease. Emerging evidence suggests has emerged that increased fruit and vegetable intake and decreased intake of saturated fats, simple sugars, and processed foods can improve cardiovascular health. Anthocyanins (color pigments) have anti-inflammatory and antioxidant capacities but are of low bioavailability. In this systematic review and metaanalysis, we investigate the possible beneficial effects of the intake of berries high in anthocyanins on MetS risk factors. We also investigate the influences of high-density lipoprotein (HDL), low- density lipoprotein (LDL), triglycerides (TG), and total cholesterol (TC). METHODS We identified 2,274 articles from PUBMED and EMBASE following a search input designed to include studies of interest of these, 21 met inclusion criteria. RESULTS The studies showed an overall reduction in low-density lipoprotein (p=0.04). Increases in HDL were found with cranberry and freeze-dried berry intake during a 4-6-week intervention. No statistically significant findings were detected for fasting glucose, Hb1Ac, insulin levels, blood pressure, oxidized LDL (OX-LDL), BMI, and overall HDL. CONCLUSIONS We conclude from this systematic review and meta-analysis that increased berry intake improves MetS key risk factors and reduces the risk of cardiovascular disease. Pronounced effects were apparent for concentrated berry products, such as freeze-dried strawberries.


Cholesterol, cardiovascular disease, and metabolic syndrome
ollowing the current definition of metabolic syndrome (MetS) introduced in 1998, it has been estimated that more than one-fourth of the adult population worldwide is affected by MetS [1,2]. Common pharmaceuticals are currently used to treat the symptoms of MetS, including blood pressure regulators, cholesterol-lowering medications and drugs to treat type 2 diabetes mellitus (T2DM). MetS patients are recommended lifestyle changes, including increased exercise, increased intake of fruits and vegetables, and decreased consumption of simple sugars, foods high in saturated fatty acids, and processed foods [3]. Studies show that these changes cause effective improvements in their condition [4].
Several studies (including both meta-analyses and observational studies) have found correlations between increased daily intake of fruits and vegetables and improved cardiovascular health [5][6][7]. Likewise, some studies have investigated the effects of diets high in fruits on blood lipids and sensitivity to insulin and glucose and showed a decreased risk of cardiovascular disease (CVD) [5,6]. The mechanism behind this is complex, but it has been suggested that consumption of fruits high in anthocyanins may induce some of these beneficial effects [6,8,9]. In this systematic review and meta-analysis, we focus on studies investigating the effects of anthocyanin -rich berries on high-density lipoprotein (HDL) in patients with MetS to evaluate current evidence for beneficial effects. The possible beneficial effects on low-density lipoprotein (LDL) also has been evaluated in these studies. Through this systematic review and meta-analysis, we provide an overview of the latest evidence from randomized controlled trials (RCTs) evaluating anthocyanin-rich berries.

Dyslipidemia and cardiovascular disease
MetS patients are more likely to develop CVD than non-obese and non-MetS patients [10,11]. The underlying mechanisms are complex, but common to patients with CVD and MetS are decreased levels of HDL and increased levels of both LDL and TG [12,13]. MetS patients have more total body fat and also more visceral fat than normal healthy individuals [14]. This can lead to increased inflammation, in particular, a chronic state of low-grade inflammation [15]. A 2019 study showed that visceral fat, in particular, makes a greater contribution to chronic inflammation per volume than any other type of fat storage [16].
The underlying mechanism causing inflammation is not fully understood, but two mechanisms have been proposed. The first mechanism was illustrated in studies by Alvehuset al. [17] and Cao et al. [18]. The studies showed that the increase in the transcription of CC chemokine receptor 2 (CCR2), macrophage migration inhibitory factor (MIF), and tumor necrosis factor α (TNF-α), a chemokine receptor, inflammatory cytokine, and pro-inflammatory factor, was greater in visceral adipose tissue than other types of adipose storage, leading to increased inflammation [17,18]. A 2007 study by Fontana et al. demonstrated that the portal vein in severely obese individuals had a 50% higher concentration of interleukin 6 (IL-6), a proinflammatory factor, than the radial artery, suggesting that visceral fat has a significant effect on introducing chronic inflammation in obese and MetS patients [19].
Regardless of the mechanism, the effect of visceral fat always seems to be an increased level of circulating TNF-α and IL-6, two pro-inflammatory factors [20]. These factors may activate several inflammatory pathways to cause systematic inflammation and increased secretion of insulin which may result in lipid abnormalities, including lowered HDL and raised LDL and TG [21,22]. Therefore, it is essential to investigate how to improve blood lipids and to lower the inflammatory state of MetS patients to prevent CVD.

Blood lipids and cardiovascular disease
Hyperinsulinemia leads to decreased HDL and increased TG. Increased TG levels have been shown to result in more TG binding to HDL, and in turn, this decreases both HDL concentration and reverse cholesterol transport. The reason for this event is that hepatic lipases have increased affinity for TG-rich HDL particles, i.e., breaking down HDL, which results in further decreased HDL levels [23,24]. The resulting increase in TG release from HDL may cause increased risk of CVD [12].
MetS patients often have elevated LDL levels and chronic low-grade inflammation, but are also likely to have elevated and unwanted production of reactive oxygen species (ROS) [15,25,26]. ROS is one of the main factors responsible for oxidizing LDL to result inoxidized LDL (OX-LDL) [27]. OX-LDL can bind to several non-native receptors on different endothelial cells including lectin-like oxidized LDL receptor 1 (LOX-1). Binding of OX-LDL to LOX-1 induces monocyte adhesion to endothelial cells, a pre-requisite for macrophage foam cell formation in atherosclerosis, and positive feedback on the regulation of LOX-1 resulting in more LOX-1 receptors [28] (Figure 1). This process causes an upregulation of the transcription of adhesion molecules, including intercellular adhesion molecule1 (ICAM-1), vascular adhesion molecule1 (VCAM-1), and P-selectin [29].
Elevated TGs affect the endothelial cells of the blood vessels in several ways. Flow-mediated dilation studies evaluating endothelial cell function have shown that elevated TG levels resulted in reduced arterial dilatation [36]. The same study concluded that chronically elevated levels of TG in hyperinsulinemic patients may lead to increased oxidative stress.
Also, postprandial hypertriglyceridemia has been shown to increase the expression of aleukocyte receptor known as lymphocyte function-associated antigen 1 (LFA-1), an essential integrin involved in the recruitment of immune cells ininflamed tissue. This mayresult in further elevated postprandial inflammation, imposing additional risk of CVD on the patient [37].
Furthermore, TG intake has been shown to affect hormonal regulation, resulting in elevated resistin levels, for example [38]. Resistin has been shown to increase transcriptional levels of pro-inflammatory cytokines, including TNF-α, interleukin1 (IL-1), IL-6, and IL-12. We know TNF-α and IL-6 can decrease HDL levels [39], and IL-12 is a proinflammatory cytokine that further elevates the inflammatory state.
In summary, dyslipidemia has devastating effects on health by increasing CVD risk, as discussed above. Therefore, it seems essential to improve the lipid profile of dyslipidemic MetS patients, a patient group that is constantly increasing in number. Global interest in ameliorating and possibly preventing CVD nonpharmaceutically has been sparked, and anthocyanins have received a great deal of interest as a potential CVD treatment.

Anthocyanins and their effects on high-density lipoproteins
A 2014 study by Yanna Zhu et al. showed how 320 mg/d of anthocyanins derived from bilberries and black currants resulted in a significant increase in HDL in hypercholesterolemic subjects following a 24week intervention period [9]. The mechanism behind the improved HDL concentrations has not been fully elucidated. The authors suggest that anthocyanins or their derivatives may lead to the activation of serum paraoxonase and arylesterase 1 (PON1), which prevents HDL from being oxidized and broken down, thus increasing HDL's cholesterol efflux capacity allowing an efficient reverse cholesterol transport [9].
Also, a 2015 in vivo study using mice by Nicholas Farrell et al. showed how mice fed an anthocyanin-rich diet had no significant serum HDL changes, but showed gene expression alterations that were associated with improved liver and HDL functions [40]. Hepatic and intestinal mRNA changes in mice fed anthocyanin-rich diets had increased transcription of ApoA1, PON1, serum amyloid A1, lecithin cholesterol acyltransferase, and apolipoprotein J, all of which led to improved HDL function. Also, a reduction in serum chemokines and ligand 2 (CCL2) was seen, which represented a lowered state of inflammation.
In a 2013 study by Kianbakht et al. much lower anthocyanin doses were used resulting in significantly increased HDL serum concentrations [41]. In this study, a total of only 9.8 mg/d of whortleberry-extracted anthocyanins were given daily for two months. The intervention group showed a 37.5% increase in serum HDL in the post-intervention period relative to baseline values.

Anthocyanin and their effects on low-density lipoproteins
An RCT from 2014 conducted by Rasool Soltani et al. showed a significant decrease in LDL levels in hypercholesterolemic subjects [42]. The intervention group (n=25) consumed a 45 mg whortleberry anthocyanin capsule twice daily for 4 weeks. The LDL serum levels were lowered by a mean of 11.44±3.28 mg/ dl.
Likewise, a 2008 study by Lee et al. on persons with T2DM with elevated LDL (mean LDL of 127.61±29.93 mg/dl) showed a significant LDL decrease following a 12-week intervention period [43]. The intervention group received 500 mg cranberry extract 3 times daily for a period of 12 weeks. After intervention, their LDL levels had decreased to a mean concentration of 112.14±29.93 mg/dl. The exact anthocyanin dose was not determined, but cranberries are regarded as being among the berries with the highest anthocyanin content [44].
To date, no definitive mechanism has been determined as to how anthocyanin consumption can result in lowered LDL levels. However, a 2009 study suggests that anthocyanins can impede cholesteryl ester transfer protein (CETP), a deactivation that is associated with decreased formation of both LDL, very low-density lipoproteins (VLDL), and intermediatedensity lipoproteins (IDL) [8].
A 2018 cellular study by Kimet al. showed that Caco-2 cells treated with anthocyanin-rich blackcurrant extract showed an upregulation of the low-density lipoprotein receptor, ultimately lowering LDL. It is believed that a similar effect occurs in humans [45]. on whole-body insulin sensitivity [52]. In this study, 32 obese, non-diabetic, insulin-resistant subjects were recruited, and a significant increase in insulin sensitivity was seen in the intervention group. No mechanism for the increase in insulin sensitivity has yet been determined in human subjects, but a mechanism has been proposed in mice, where bilberry anthocyanins were reported to have activated adenosine monophosphate-activated protein kinase (AMPK) in mouse skeletal muscles, liver, and white adipose tissue [53]. This activation apparently caused an upregulation of glucose transporter four in skeletal muscles and white adipose tissue, while down regulating glucose production in the liver.

Anthocyanins and their antioxidant capacities
Only a few studies evaluate berry antioxidants in humans, but these show promising results. A 2015 clinical trial by Sergio Davinelli et al. showed that 486 mg of anthocyanins daily significantly reduced levels of OX-LDL post intervention [54]. However, following a literature search, it is evident that there are no studies demonstrating that the anthocyanins administered prior to their metabolism in humans are responsible for the postulated antioxidant function of anthocyanins . Therefore, it might be their

Anthocyanins and their effects on triglycerides
In 2015, Pei-Wen Zhang et al. showed that TG serum levels significantly lowered by as muchas 17.27% after 12 weeks of blueberry consumption. The underlying mechanism is not fully understood, but a potential mechanism was proposed by Dan Li et al. in 2015. They suggested that anthocyanins may result in a reduction in serum apolipoprotein-B (ApoA-B) and apolipoprotein-C (ApoA-C) containing TG-rich particles [46].

Anthocyanins and their effects on insulin secretion and sensitivity
In 2010, April J. Stull

Search strategy and study selection
This review includes articles identified by systematic search as well as other articles identified by the research team outside the search strategy. These articles include studies carried out in vitro prior to human studies to identify possible benefits of anthocyanins. The meta-analysis includes only articles identified by the systematic search strategy. In this meta-analysis, both PUBMED and EMBASE have been searched systematically for randomized clinical trials up to April 31, 2020. Appendix 1 shows the terms used for PUBMED and EMBASE searches. The study was registered on Prospero with ID CRD42020181167.
Human participants with diagnosed MetS (MetS) or fulfilling 3 out of 5 diagnostic criteria for MetS diagnosis were included in the meta-analysis [1]. The criteria for MetS diagnosis were applied according to the International Diabetes Federation Guidelines, updated in 2020 [56]: --Waist circumference >80 cm --TG level >150 mg/dl --HDL <40/50 mg/dl --Blood pressure >130/85 --Fasting glucose >100 mg/dl Despite BMI not being required for MetS diagnosis, BMI above 25 has shown similar systemic consequences as MetS including the development of T2DM [57]. Likewise, if the articles did not meet the criteria, we estimated that if participants had BMI >27, their waist circumference would be above 80 cm.
The following inclusion criteria also had to be met: --Human studies --Adults ≥18 years of age --Intervention length ≥4 weeks --Reported mean end values at the end of the intervention with their corresponding SD, SEM or confidence intervals (CI) --The intervention products were provided by the study personnel to the participants --Publications dating from 1970 or more recently --Intervention products included fresh, freezedried, juices and extracts from berries with proven anthocyanin content. The exclusion criteria were as follows: - --Intervention length < 4 weeks --Interventions consisting of recommendations, telephone calls, etc.

--Reviews
--Observational studies --Meta-analyses --Data missing or not presented All articles were evaluated based on title, abstract, and full text by two independent reviewers. The outcome of interest in this systematic review and metaanalysis was HDL level. However, an evaluation of LDL and TC as well as OX-LDL, BMI, blood pressure, fasting glucose, Hb1Ac, and insulin levels also had been carried out based on the included studies. If the publication appeared prior to 1970, the study was not included.

Quality assessment
Two authors (Wilkins and Christiansen) independently assessed all articles following the PUBMED and EMBASE search. All included articles went through a risk of bias assessment by means of the Cochrane collaboration's tool that includes the following criteria: --Random sequence generation --Allocation of concealment --Blinding of participants and personnel --Blinding of outcome assessment --Incomplete outcome data --Other possible causes for bias In the case of divergence between the assessments by the two authors, a third author (Lambert) independently assessed the articles to reach consensus. The quality of included studies was evaluated using the risk of bias assessment tool from the Cochrane Handbook for Systematic Reviews of Interventions (version 5.3.0). It includes tools for selection bias (random sequence generation and allocation concealment), performance bias (blinding of participants and personnel), detection bias (blinding of outcome data), attribution bias (incomplete outcome data), reporting bias (selective reporting), and other sources of bias.

Data extraction
Only studies publishing mean ±SD, mean ±SEM, or a 95% CI for endpoint values were used. We used Cochrane Handbook version 5.1.0 section 7.7.3.2 to convert mean ±SEM and CI to mean ±SD. All HDL, LDL, TC, and other values were converted to mg/dl. Fasting glucose, Hb1Ac, insulin levels, and BMI were converted to mg/dl, percent, mU/l, and kg/m 2 , respectively. Also, subgroup analyses were performed for BMI >30, HDL levels, anthocyanin doses, intervention length, and kind of berry used for intervention.

Statistical analysis
Mean ±SD values for end values within both the intervention and the control group were all pooled in Review Manager (RevMan) 5.3 (the Cochrane Collaboration's software) to analyze and evaluate the effects on HDL, as well as other parameters, in both the control and intervention group following the intervention period. RevMan includes tools for protocol and review preparation, assessing characteristics of studies, and comparison tables. It has tools to perform meta-analyses and can present results graphically.
With regard to one study (P. J. Curtis et al. 2019), the 95% CI for HDL concentrations post intervention with one cup of blueberries measured in mmol/l was not presented with adequate decimals to calculate a ±SD of more than 0 following a conversion from mmol/l to mg/dl [58]. The 95% CI presented was (1.2, 1.2). To ensure a ±SD for the analysis, the CI used was set to (1.21, 1.22) (Appendix A2).
To evaluate and determine heterogeneity, we used the measure of inconsistency, where I 2 >50% was considered as substantial heterogeneous. We used the random effect model to determine statistical significance. Weighted mean difference and 95% CI were obtained from the pooled studies and visualized in forest plots to illustrate the obtained data. With regards to publication bias, the shape of the funnel plot indicated a certain asymmetry between the differences in HDL levels when anthocyanin-rich berry administration was compared with placebo. This appears more pronounced in smaller studies and becomes less prominent with larger scale trials which seemed to be more robust. It indicates the presence of systematic differences between smaller and larger studies, with a higher possibility of bias in smaller studies. However, publication bias was not confirmed by Egger´s test (p=0.47) (Appendix A3).

Selection of studies
We identified 2,274 articles by searching PUBMED and EMBASE, 21 of which have been included in this meta-analysis. After removing duplicates, screening of abstracts, and screening of full texts, 2,253 articles were discarded leaving a total of 21 articles to be included in the study (Figure 2). Six 6 studies did not include endpoint data for evaluated parameters, but authors were contacted; one author responded and provided the data, the remaining 5 have been excluded.
All 21 studies were randomized controlled trials or randomized crossover studies. A total of 25 interventions provided a total of 1,355 participants in this meta-analysis. Studies included participants with dyslipidemia [41,59,68,69,[60][61][62][63][64][65][66][67], MetS [58,70,71], T2DM [46,[72][73][74][75], and nonalcoholic fatty liver disease (NAFLD) [49]. All participants in this analysis who had not already been diagnosed with MetS were evaluated were administered in the studies. The intervention duration varied from four weeks to six months. Prior to intervention, we observed no statistically significant differences between the placebo group and intervention group in all studies. All studies included in the meta-analysis were evaluated for risk of bias (Figure 3). The differences between the studies were as follows: --Ten studies had no information on the method of randomization [46,58, ]. There were also other causes for possible bias in the following studies: --One study had 45 participants in the control group, but only showed data on three [74]. --One study did not provide the amount of anthocyanin in their product; so this was estimated [66].In one study, there was a significant difference in HDL levels between the intervention and placebo group in the preintervention period [60] ( Table 2).

Effects of berries high in anthocyanins on HDL
A comparison between all 1355 participants from the 21 studies and 25 interventions was made. HDL endpoint values from the intervention group were compared with the HDL endpoint values from the control group. To illustrate the comparison, we created a forest plot (Figure 4). There was no statistical significance, but we observed a tendency towards a beneficial effect of HDL concentrations following the intervention period (p=0.14, 95% CI -0.33, 2.37, and I 2 of 89% (p<0.00001) showing significant heterogeneity).
We performed a subgroup analysis on cranberry interventions and the effects on HDL. We obtained a statistically significant result from three studies, five interventions, and 267 participants. After intervention, HDL levels for the intervention group were significantly higher, namely 2.01 mg/dl higher than in the placebo group (p=0.007, 95% CI 0.55, 3.48, and I 2 of 87%) suggesting that cranberries can increase HDL in MetS patients (Figure 5). Relative to other berries high in anthocyanins, cranberries may contain more readily bioavailable anthocyanins than other berries, which may have increased HDL in studies using cranberries to a greater degree than in those using other fruits [76].
Following a categorization of intervention length, four studies with a total of six interventions and 281 participants showed an increase in HDL levels for the intervention group following intervention duration of 4-6 weeks. A slight increase of 1.78 mg/dl was observed with a p-value of 0.03 and 95% CI0.22, 3.34 with I 2 = 90% (Figure 6). No effect was seen after an intervention shorter than 4-6 weeks. No studies lasting k of bias summary of inc for MetS; they were regarded as MetS patients if they fulfilled three out of five criteria.
Different types of intervention and placebo Legend: Age given with ±SD, SEM, or range if provided. If age was not provided in articles per group for intervention and control groups, age shown is for both groups and sexes combined. Abbreviations: Inv -intervention group, Con -control group, LD-FD -low-dose freeze-dried strawberries, HD-FDS -high-dose freeze-dried strawberries, LD-C -low-dose freeze-dried strawberry control group, HD-C -high-dose freeze-dried strawberry control group, RC -randomized crossover, RCT -randomized controlled trial, T2DM -type 2 diabetes mellitus.
for more than six weeks showed any increase, but this may be due to poor adherence of participants, i.e., participants becoming less adherent to consuming the provided anthocyanin products after an extended time period.

Effects of berries high in anthocyanins on LDL
We also carried out a comparison between the intervention and control group regarding LDL end values. All 21 studies and 25 interventions were compared by creating a forest plot (Figure 7), showing significantly lower LDL levels in the intervention compared to the control group (p=0.04). The mean difference was -3.21 mg/dl with a 95% CI -6.31, -0.12, and I 2 of 46% (p=0.006), showing significant heterogeneity.
We performed a subgroup analysis on freezedried strawberry interventions and their effects on LDL levels. We found a statistically significant mean difference of -11.06 mg/dl in the intervention group after intervention (p= 0.002, 95% CI-18.00, -4.12) (Figure 8).

Effects of berries high in anthocyanins on TC
A forest plot was also created from the TC end values from the 21 studies and 25 interventions (Figure 9). The forest plot showed no statistically significant differences between end values for the intervention and the control group (p=0.08). However, a tendency towards lower TC values was seen in the intervention group showing a mean difference of -3.64 mg/dl with a 95% CI -7.77, 0.49, I 2 of 57%, and p=0.0003, showing significant heterogeneity.
Four studies and five interventions investigated freeze-dried strawberries and showed a decreasing effect on TC in the intervention group with a similar p-value of <0.00001. The mean difference between the placebo and intervention group was -12.20 mg/dl (95% CI-15.65, -8.75, and I 2 of 0%) (Figure 10). Table 3 contains the results of the subgroup analyses.

Discussion
In this systematic review and meta-analysis, we evaluated the effects of interventions with berries high in anthocyanins for their ability to improve blood lipid profiles in MetS patients. HDL, LDL, and TC were evaluated as well as other parameters which are abnormal in MetS patients, including fasting glucose, Hb1Ac, insulin levels, blood pressure, and OX-LDL [77].
The meta-analysis showed a statistically significant decrease in LDL concentrations in the pooled data from all intervention groups compared with placebo (p=0.04). This finding correlates with the conclusions drawn by L. Yang et al. 2017, who reported a significant improvement in LDL [78]. However, the mean change in the present analysis was only 3.21 mg/dl and cannot be regarded as clinically relevant with regards to cardiovascular health [79]. Significant heterogeneity was present in the pooled intervention group, suggesting a significant variation between the studies. No other statistically significant differences were found for the 21 pooled studies (for HDL or TC).
Regarding fasting glucose, Hb1Ac, insulin levels, blood pressure, OX-LDL, and BMI end values, the only statistically significant difference seen was for BMI (-0.69 kg/m 2 , p=0.02, for the intervention group). A mean decrease of -0.69 kg/m 2 in BMI following an intervention may not be clinically relevant in terms of cardiovascular health. Also, as only 8 interventions were included, a more conclusive result may be obtained if more studies were available (I 2 = 0%).
Fasting glucose and diastolic blood pressure remained unchanged, with p-values of 0.09 and 0.06 respectively. Statistically significant results were observed regarding HDL, LDL, and TC following an intervention with cranberries or for interventions   lasting between four to six weeks; improvements of 2.01 mg/dl and 1.78 mg/dl were seen in the intervention groups, respectively (p=0.007 and p=0.03, respectively). An increase of one mg/dl in HDL has been shown to reduce the risk of CVD by 2% -3% [80,81]. We found a statistically significant result regarding LDL levels when pooling all studies with freeze-dried strawberry interventions. In this case, LDL levels decreased in the intervention group by a mean of 11.04 mg/dl, p=0.002. It has been shown that decreasing LDL by 38.68 mg/dl reduces the risk of CHD by 23%. Hence, a reduction of 11.04 mg/dl, as observed in the present analysis, is likely to exert substantial improvements on CVD risk in MetS patients [82]. However, it is also essential to evaluate the ratio of LDL/HDL as elevated serum LDL/HDL ratio was independently associated with increased risk of CVD [83].
Finally, we found a statistically significant difference in TC levels after intervention with freeze-dried strawberries, with lower levels in the intervention group. The mean difference was -12.20 mg/dl (p<0.00001). In this regard, freeze-drying is a powerful method of increasing the concentration of nutrients and bioactive compounds by removing water from the berries. This  Figure 6. Forest plot showing the effects of intervention length of 4-6 weeks on HDL. Data are presented as mean difference and 95% CI using a random-effect model.  Figure 9. Forest plot showing the effects of the intervention included in the meta-analysis on TC levels. Data are presented as mean difference and 95% CI using a random-effect model.  Figure 10. Forest plot showing the effects of freeze-dried strawberry interventions on TC levels. Data are presented as mean difference and 95% CI using a random-effect model. may have had a significant effect on the results. There are common limitations to meta-analyses when working with dietary interventions that must be considered when interpreting the results. Small sample sizes can cause I 2 values to be low, even when studies are different. This particularly applies to subgroup analyses in the present study, where sample sizes were small indicating large heterogeneity. The large heterogeneity observed in some of these analyses maybe the consequence of different interventions, study designs, sample sizes, durations, and groups of participants.
Another major limitation to dietary interventions is the placebo product. It is of great importance to produce a placebo that is of similar appearance, texture, and flavor as the intervention product. A great variety of natural flavors and colorings are derived from fruits, especially fruits that are high in anthocyanins. Unless the studies have analyzed the placebo product with regards to anthocyanins, it is possible that the placebo could produce unwanted effects, although considering the low concentrations used for these purposes it must be regarded as negligible.
The time of the year also can influence findings; seasonal changes in dietary habits may have affected the outcome of the interventions [84]. Thus, it is not possible to attribute any of the tendencies or changes to anthocyanins alone. Anthocyanin-rich berries are also rich in vitamins and nutrients, carrying several bioactive components that potentially improve cardiovascular health in a synergistic way [85,86]. Vitamin C, vitamin E, and carotenoids are all found in berries and are known for their antioxidant capacities [87]. These bioactive compounds may act in a synergistic way with anthocyanins to achieve the positive results. Therefore, the observed results should be interpreted with caution. Moreover, there are also differences in the composition of bioactive substances between different varieties of berries, the conditions in which they are cultivated, and climate conditions, which may have introduced another bias into the meta-analysis. We cannot exclude the possibility that other bioactive compounds (e.g., fiber type and content) explain some of the beneficial effects detected by this meta-analysis.
Along similar lines, studies show that anthocyanin bioavailability is as low as 1%, suggesting that the observed beneficial effects cannot be attributed to anthocyanins, but to other bioactive components in the berries [88]. Likewise, some studies suggest that ingested anthocyanins are converted to bioavailable polyphenols by the gut microbiota, which may contribute to the positive effects found in the present meta-analysis [89][90][91].
Finally, vegetables and fruits other than those included in the studies also contain anthocyanins. Hence, the participants, including those in the placebo group, would have ingested anthocyanins from other sources. Therefore, it is difficult to say whether the anthocyanin in question was obtained from the interventions or from their usual daily diet. It would have been beneficial if all placebo groups had not been allowed to eat anthocyanin-containing food items to strengthen the results associated with anthocyanins in the diet.
Also, MetS is more common in elderly than in young adults and general nutrient bioavailability is poorer in the elderly than in young. Therefore, it would have been beneficial to have studies that discriminate on age so that it would be possible to determine how effective anthocyanins are relative to age [92].
The strength of this meta-analysis is the broad search for articles using MESH and EM-Tree terms. Moreover, none of the interventions in the metaanalysis had side effects or caused dangerously high levels of oxidative stress (measured by OX-LDL levels), blood lipids, or other parameters investigated.
A future perspective maybe to investigate anthocyanins in a more exclusive environment, i.e., only anthocyanin extracts should be investigated. This would enable a more robust result with regards to anthocyanins' health benefits and exclude synergistic effects occurring with food matrices. Another approach would be to increase both the bioavailability and concentration of anthocyanins and other phenolic compounds in berries or berry products. This was shown by X. Du and A. D. Myracle in 2018 who changed the anthocyanin profile and raised the antioxidant capacity of aronia berries though kefir fermentation [93]. Therefore, it might be of interest to ferment berries or berry products and conduct an intervention with these products to evaluate their effects on the presently investigated parameters.
Another future perspective may be to use freezedried berries because promising in vivo results have been observed. A 2013 study by L. Brader et al. showed that following ingestion of 5 g/d freeze-dried bilberry powder for 8 weeks, a 60% reduction in LDL levels was seen in diabetic rats [94]. However, these results must be interpreted with caution as rodents in general have a different lipid metabolism than humans [95]. However, it still might be worth evaluating similar studies in humans with T2DM through a meta-analysis.

Conclusions
In this systematic review and meta-analysis, we investigated the effects of berries high in anthocyanins and their ability to improve the cardiovascular health of MetS patients. The overall analysis with all 25 interventions and more than 1300 participants showed no statistically significant rise in HDL levels. However, an increase was found in the cranberry intervention, and when the intervention period lasted 4-6 weeks. We found a statistically significant reduction in LDL in the intervention group. We found no statistically significant decrease in TC for pooled interventions.
The freeze-dried strawberry interventions also resulted in a significant decrease in TC and LDL. Hence, the data in this meta-analysis showed correlations between anthocyanin consumption and improvements in certain lipid biomarkers related to MetS. More data and studies are needed to develop recommendations regarding consumption of anthocyanin-rich berries for MetS, as effects from other bioactive compounds found in the berries may have interfered with the results in the present analysis.

Funding: None
Conflict of interest statement: The authors report no conflict of interests.
Author contributions: These are described in the text provided.