Effects of Anthocyanin Supplementation on Reduction of Obesity Criteria: A Systematic Review and Meta-Analysis of Randomized Controlled Trials

Anthocyanins, water-soluble flavonoids that produce red-to-blue pigment in plants, have antioxidant properties and have been developed as a functional food to fight obesity. In randomized controlled trials (RCTs), a systematic review with meta-analysis (SR-MA) was used to investigate these anti-obesity effects. Using search engines (PubMed, EMBASE, Cochrane-library, and CINAHL) and keywords (anthocyanins, BMI, WC, WHR, and inflammatory biomarkers), 11 out of 642 RCTs (28.3–500 mg/day of anthocyanins for 4 to 24 weeks) were included. The results showed a significant reduction in body mass index (BMI) (MD = −0.36, 95% CI = −0.58 to −0.13), but body weight (BW) and waist circumference (WC) did not change. Anthocyanins decreased BMI in the non-obese (non-OB) group in five RCTs (BMI ≤ 25; MD = −0.40 kg/m2; 95% CI = −0.64 to −0.16;) but did not affect BMI in the obese (OB) group. A subgroup analysis of six RCTs showed that fewer than 300 mg/day reduced BMI (MD = −0.37; 95% CI = −0.06 to −0.14), but ≥300 mg/day did not. A treatment duration of four weeks for four RCTs was sufficient to decrease the BMI (MD = −0.41; 95% CI = −0.66 to −0.16) as opposed to a longer treatment (6–8 or ≥12 weeks). An analysis of the effect of anthocyanins on the BMI showed a significant fall among those from the Middle East compared to those from Asia, Europe, South America, or Oceania. In conclusion, the anthocyanin supplementation of 300 mg/day or less for four weeks was sufficient to reduce the BMI and BW compared to the higher-dose and longer-treatment RCTs. However, further studies might be conducted regarding the dose- or period-dependent responses on various obese biomarkers.


Introduction
Anthocyanin is a natural soluble pigment in the flavonoid group. Approximately 635 or more structures have been found in nature and more than 30 types of anthocyanins and anthocyanidins have been identified based on the number and position of the hydroxyl group [1,2]. The common aglycones are pelargonidin (Pg), cyanidin (Cy), peonidin (Pn), delphinidin (Dp), petunidin (Pt), and malvidin (Mv), but Cy-3-glucoside was widely distributed [3,4]. Anthocyanins are mostly absorbed through the gastric wall with absorption rates of 10-22%, depending on chemical structure, and the bioavailability is approximately 0.26-1.8% [5][6][7]. Anthocyanins produce antimicrobial, antioxidant, and anti-inflammatory effects and play a role in the prevention and treatment of numerous chronic conditions,

Inclusion and Exclusion Criteria
This meta-analysis based on the PICOS design included (a) randomized controlled trials (RCTs) with anthocyanin supplements in the control group that did not receive the intervention; (b) men and women over 18; and (c) changes in BW, BMI, WHR, leptin, and adiponectin. The exclusion criteria were nonhuman subjects; non-intervention; intervention in anthocyanin deficiency; anthocyanins combined with other medications; non-RCT studies (cohort, case-control, or cross-section); review or commentary; discordance in the outcome of MetSyn; or non-English publication.

Quality Assessment (ROB)
The reliability of the articles in the final set of the meta-analysis was evaluated through quality assessment using the Cochrane Collaboration Risk of Bias (RoB), which combines a checklist and domain evaluation format with simplified questions. It was developed to minimize the possibility of subjective or arbitrary responses to identical questions as well as variation in assessment results due to differences in the understanding and proficiency of the reviewers concerning the methodology. The quality assessment was performed by two independent reviewers (S.P. and M.C.), while the level of RoB was determined through discussion. The third independent reviewer (M.L.) resolved any disagreements in RoB evaluation. The following seven items were examined based on the risk levels "high", "low", and "unclear": Random sequence generation, allocation concealment, blinding of the participants and personnel, blinding of the outcome assessments, incomplete outcome data, selective reporting, and other sources of bias.

Data Analysis
The mean differences (MD) between the baseline and final value of the factors were extracted from the studies. A statistical meta-analysis was conducted using RevMan software (Version 5.3: Review Manager [RevMan], Nordic Cochrane Centre, Copenhagen, Denmark). The effect size was a quantitative measure that summarized the results, reflecting the magnitude or intensity of the relationship among the different study parameters. In this review, the effect size was based on differences in mean values as the treatment continued. The confidence interval (CI) for significance testing was 95%. For the included primary articles that did not report standard deviation (SD), the method of calculation reported by Higgins et al. was used [24]. The result of the meta-analysis was presented on forest and funnel plots [25]. The forest plot was presented in line with the statistical analysis to show the effect size (quadrangle) and 95% CI (line of symmetry). A fall in the width of the 95% CI increased the area of the quadrangle. Statistical analysis included the significance (p value), summary estimate (mean ± SD, 95% CI), and heterogeneity (I 2 ). When the standard deviation (SD) for mean differences was not reported, it was calculated by the following formula: SD = square root ((SD baseline) 2 + (SD final) 2 − (2R × SD baseline × SD final)), assuming the correlation coefficient (R) = 0.5 [26]. Funnel plots were used to assess publication bias. The Higgin's I 2 value was used as a statistical test to verify heterogeneity [24]. In general, 0% ≤ I 2 ≤ 25% indicated a low level of heterogeneity; 25% ≤ I 2 ≤75% indicated a moderate level, and 75% ≤ I 2 ≤ 100% indicated a considerably high level. Additionally, the sub-group analysis was estimated using the same methods.

Study Selection
The search identified 1379 articles, and after the removal of duplicates, 737 were screened by title and abstract. Most (n = 620) were excluded because they were not relevant: Duplicates (n = 22), nonrandomized controlled trial (n = 53), and review or non-human study (n = 545). After assessing the full text of 22 potentially related articles, eleven articles were included. The most important reasons for exclusion were non-compliance with SR-MA criteria: No placebo group (4); subjects unrelated to the aim of the review (athletes, other functionality, age) (4); or sex bias, a short communication study, or lacked anthocyanin information (3). Figure 1 shows the study identification and selection process.

Characteristics of the Included RCTs
Among the 11 RCTs selected for analysis (n = 833), nine (n = 401), were included for the analysis of the correlation between BMI and anthocyanin supplementation, five (n = 233) for the correlation between BW and anthocyanin supplementation, and four (n = 148) for the correlation between WC and anthocyanin supplementation (Table 1). However, no study reported the correlation between inflammatory biomarkers and anthocyanin supplementation; hence, this was not considered in this review. The number of subjects in the 11 RCTs ranged between 16 and 80; the daily dose of anthocyanin supplementation ranged between 28.3 and 500 mg/day, and the duration of intervention ranged between 4 and 24 weeks. The countries where these were conducted were China [11], Serbia [12], Italy [14], South Korea [15], Germany [16], Italy [19], the U.K. [20], Denmark [21], Brazil [22], Australia [23], and Iran [27]. The analysis of subject characteristics showed that four RCTs did not target patients [16,[19][20][21]; one RCT targeted patients with type II DM [11]; four RCTs targeted overweight or obese patients [12,14,22,23]; one RCT targeted patients with CVD [14]; and one RCT targeted patients with hyperlipidemia [27].

Characteristics of the Included RCTs
Among the 11 RCTs selected for analysis (n = 833), nine (n = 401), were included for the analysis of the correlation between BMI and anthocyanin supplementation, five (n = 233) for the correlation between BW and anthocyanin supplementation, and four (n = 148) for the correlation between WC and anthocyanin supplementation (Table 1). However, no study reported the correlation between inflammatory biomarkers and anthocyanin supplementation; hence, this was not considered in this review. The number of subjects in the 11 RCTs ranged between 16 and 80; the daily dose of anthocyanin supplementation ranged between 28.3 and 500 mg/day, and the duration of intervention ranged between 4 and 24 weeks. The countries where these were conducted were China [11], Serbia [12], Italy [14], South Korea [15], Germany [16], Italy [19], the U.K. [20], Denmark [21], Brazil [22], Australia [23], and Iran [27]. The analysis of subject characteristics showed that four RCTs did not target patients [16,[19][20][21]; one RCT targeted patients with type II DM [11]; four RCTs targeted overweight or obese patients [12,14,22,23]; one RCT targeted patients with CVD [14]; and one RCT targeted patients with hyperlipidemia [27].

Anthocyanin Supplements on the Reduction of BW or WC
The analysis of the effect of anthocyanin supplementation on BW in five RCTs (n = 233) showed that it increased by 0.15 kg, compared with that of the control group (95% CI = 0.01-0.29; I 2 = 0%; p = 0.04) ( Figure 4A). The funnel plot showed that the five RCTs were in symmetry, indicating a low publication bias. After categorizing the subjects into healthy (four RCTs, non-OB; BMI < 25) and obese (one RCT, OB; BMI ≥ 25) individuals, the BW was higher by 0.15 kg in the former after anthocyanin supplementation (95% CI = 0.01-

Anthocyanin Supplements on the Reduction of BW or WC
The analysis of the effect of anthocyanin supplementation on BW in five RCTs (n = 233) showed that it increased by 0.15 kg, compared with that of the control group (95% CI = 0.01-0.29; I 2 = 0%; p = 0.04) ( Figure 4A). The funnel plot showed that the five RCTs were in symmetry, indicating a low publication bias. After categorizing the subjects into healthy (four RCTs, non-OB; BMI < 25) and obese (one RCT, OB; BMI ≥ 25) individuals, the BW was higher by 0.15 kg in the former after anthocyanin supplementation (95% CI = 0.01-0.29; I 2 = 0%; p = 0.03) ( Figure 4B). When the subjects were categorized into two groups according to dosage, the BMI was higher by 0.15 kg in Group 1 (four RCTs; ≤300 mg/day) than in Group 2 (one RCT; >300 mg/day; 95% CI = 0.01-0.29; I 2 = 0%; p = 0.04) ( Figure 4C). When the subjects were categorized into two groups according to intervention duration, the BW in Group 1 (two RCTs; ≤4 weeks) was significantly higher than that in Group 2 (three RCTs; >4 weeks), and after anthocyanin supplementation, the BW was higher by 0.15 kg in the former (95% CI = 0.01-0.29; I 2 = 0%; p = 0.04, Figure 4D).
Nutrients 2021, 13, x FOR PEER REVIEW 8 of 14 0.29; I 2 = 0%; p = 0.03) ( Figure 4B). When the subjects were categorized into two groups according to dosage, the BMI was higher by 0.15 kg in Group 1 (four RCTs; ≤300 mg/day) than in Group 2 (one RCT; >300 mg/day; 95% CI = 0.01-0.29; I 2 = 0%; p = 0.04) ( Figure 4C). When the subjects were categorized into two groups according to intervention duration, the BW in Group 1 (two RCTs; ≤4 weeks) was significantly higher than that in Group 2 (three RCTs; >4 weeks), and after anthocyanin supplementation, the BW was higher by 0.15 kg in the former (95% CI = 0.01-0.29; I 2 = 0%; p = 0.04, Figure 4D). The effect of anthocyanin supplementation on WC in four RCTs (n = 148) showed no difference to that of the placebo group (p = 0.49). The funnel plot shows that the four RCTs were in symmetry, which indicated low publication bias.

Anthocyanins Supplements on BMI and BW by National Origin
The subjects were categorized into five groups according to geography: Group 1 (four RCTs, Europe), Group 2 (two RCTs, Asia (China, South Korea)), Group 3 (one RCT, Middle East (Iran)), Group 4 (one RCT, South America (Brazil)), and Group 5 (one RCT; Oceania (Australia)). The analysis of the effect on BMI showed a significant fall in Group 3 ( Figure 5A). When the subjects were categorized into two group, Group 1 (one RCT; Asian, South Korea) and Group 2 (four RCTs; Caucasian, Europe), anthocyanin supplementation significantly increased the BW by 0.15 kg in Group 2 (95% CI = 0.01-0.29; I 2 = 0%; p = 0.03) ( Figure 5B). In summary, this SR-MA showed that while there was a significant reduction in BMI, there was no reduction in BW or WC, and ethnic difference influenced each indicator of obesity (Table 2). The effect of anthocyanin supplementation on WC in four RCTs (n = 148) showed no difference to that of the placebo group (p = 0.49). The funnel plot shows that the four RCTs were in symmetry, which indicated low publication bias.

Anthocyanins Supplements on BMI and BW by National Origin
The subjects were categorized into five groups according to geography: Group 1 (four RCTs, Europe), Group 2 (two RCTs, Asia (China, South Korea)), Group 3 (one RCT, Middle East (Iran)), Group 4 (one RCT, South America (Brazil)), and Group 5 (one RCT; Oceania (Australia)). The analysis of the effect on BMI showed a significant fall in Group 3 ( Figure 5A). When the subjects were categorized into two group, Group 1 (one RCT; Asian, South Korea) and Group 2 (four RCTs; Caucasian, Europe), anthocyanin supplementation significantly increased the BW by 0.15 kg in Group 2 (95% CI = 0.01-0.29; I 2 = 0%; p = 0.03) ( Figure 5B). In summary, this SR-MA showed that while there was a significant reduction in BMI, there was no reduction in BW or WC, and ethnic difference influenced each indicator of obesity (Table 2)

Discussion
The previous SR-MAs for anthocyanins reported on dyslipidemia [28], vascular inflammation [29], lipid composition and inflammation [30], heart disease [31], hypertension [32][33][34], DM [35], and antioxidation [36]. Thus far, only a few have been conducted on obesity-related studies for the anti-obesity effect of a flavanol complex [36] and the weight-reducing effect of resveratrol (400-800 mg/day, 1-24 weeks) [37]. In this SR-MA, we excluded the anthocyanin complex or other variants for the anti-obesity effects of anthocyanin. A notable advantage of the SR-MA is that the risk of error is reduced, which increases the reliability as the overall sample size is increased by compiling different smallscale studies. The anti-obesity effects of various anthocyanins reviewed in this analysis are thus expected to contribute to the generalization.
When the total subjects in 11 RCTs (n = 833) were categorized into healthy individuals (five RCTs, non-OB; BMI < 25) and obese individuals (four RCTs, OB; BMI ≥ 25), the BMI of healthy individuals was significantly reduced compared to that of the OB group. In the SR-MA anti-obesity effect of flavonols (58 studies), the BMI was reduced by 0.28 kg/m 2 in the flavanol-intake group, and it was reduced in group with BMI ≥ 25 more than the control [36]. Our study found that anthocyanins had a stronger effect in healthy individuals compared to those who were obese. The risk factors of diseases might not have affected the results because CVD, hyperlipidemia, or DM patients were included in both the non-OB and OB groups. In other RCT studies, the anthocyanin supplementation in chokeberry juice (28.5 mg/day) [12], blueberries (375 mg/day) [14], and Vaccinium extract (180 mg/day) [27] for 4-6 weeks had protective effects on oxidative stress, dyslipidemia, and inflammatory obesity in CVD patients. The refined anthocyanin intake of blueberry (320 mg/day) for 24 weeks reduced dyslipidemia and insulin resistance in DM patients (n = 58) [11]. These findings indicated that anthocyanin supplementation had a beneficial effect in patients with OB, CVD, and DM. Thus, the substantial BMI reduction in the non-OB group after supplementation suggests that anthocyanins can prevent obesity.
In this SR-MA, BMI was significantly reduced in the low-dose group (seven RCTs; ≤ 300 mg/day) and short-dose period (seven RCTs; 4-8 weeks) compared to the group with high-intake (two RCTs for 300 and 500 mg/day) or longer duration (two RCTs for 12 and 24 weeks). It was hard to draw a conclusion when only two studies had doses higher that 300 mg/ day and ran longer than 12 weeks. The ethnic difference according to anthocyanin supplementation was such that the BMI was reduced in individuals from the Middle East, but since only a single RCT was assessed, the finding remains inconclusive. The effect on WC was not significant in this study, which was consistent with the results of other studies [38]. The BW also increased in the non-OB group (four RCTs; BMI < 25) that had a dose of ≥300 mg/day within 4 weeks compared to the control. A significant increase in BW was observed among Europeans, whereas the BW among Koreans decreased, but as only a single RCT was assessed, a definite conclusion could not be made.
As collectively suggested by this SR-MA, a positive effect on BMI and BW reduction was anticipated for anthocyanin supplementation of ≥300 mg/day for 4 weeks, so further studies should be conducted regarding the dosage response. It was extremely difficult to estimate the daily intake or recommended level of anthocyanins due to large differences across different countries (Korea, 3.3-95.5 mg/day; U.S., 12.5 mg/day; and Europe, 30 mg/day) and individual dietary patterns [3,7,35]. Considering the mean absorption rate of 10-20%, approximately 30 mg/day upon the intake of 300 mg/day might be assumed to have a positive effect on human metabolism. Moreover, when we applied the results of anthocyanin RCTs to those who wanted to reduce BMI, we recommended minimum levels over shorter periods to reduce any risk of obesity.
The limitations of this study are as follows: First, there was no study of the literature search on the biomarkers of obesity, so the SR-MA only analyzed the correlation between obesity indicators and anthocyanin supplementation. Second, the subjects in this study included patients with DM, obesity, dyslipidemia, and CVD, which implied a potential influence on the BMI, BW, and WC. Furthermore, the disease factor made identification of the dose-dependency effect on BMI or BW reduction more difficult. Third, although the anthocyanin studies varied according to dosage, duration, and national origin, the number of RCTs (or the number of subjects) was too low to analyze the clinical effects since there has been a general lack of study on absorption rates according to ethnicity or anthocyanin polysaccharide structures. This review primarily investigated the effect of anthocyanin supplementation on obesity indicators, not on lipid composition, which has been frequently published. This study might be included in future assessments to provide comprehensive data on the preventive effect of anthocyanin on obesity.

Conclusions
We found that the anthocyanin supplementation of 300 mg/day or less for 4 weeks was sufficient to reduce BMI and BW compared to the results from higher-dose and longertreatment RCTs. Considering the daily intake, types of supplementation, and absorption rate, further RCTs should be conducted regarding the dose-or period-dependency responses on various obese biomarkers, such as adipocytokines.
Author Contributions: All authors contributed for this research articles as following: Conceptualization, and project administration, M.L.; methodology and statistical analysis, S.P. and M.C.; writing-original draft preparation, S.P.; writing-review and editing, M.L. All authors have read and agreed to the published version of the manuscript.