How effective are anthocyanins on healthy modification of cardiometabolic risk factors: a systematic review and meta-analysis

Initially, 2,512 studies were found in a preliminary search from which 2,444 studies were excluded after removing duplicates and primary screening of titles and abstracts, another 21 studies were excluded due to following reasons:

Finally, a total of 47 studies were selected for the analysis. A flow diagram showing the selection process is presented in Fig. 1.

Fig. 1figure 1

The flowchart of the study protocol on the effects of anthocyanins intake on cardiometabolic risk factors

Characteristics and bias risk assessment of included studies

The basic characteristics of the 47 randomized clinical trials (RCTs) are listed in Table 1. There were eight studies in healthy participants [22,23,24,25,26,27,28,29], four in prediabetic adults [30,31,32,33], eight in subjects with metabolic syndrome [34,35,36,37,38,39,40,41], one in elderly with mild to moderate dementia Alzheimer [42], four in overweight/obese subjects [43,44,45,46], one in adults with cardiovascular risk factors [47], seven in patients with T2D [48,49,50,51,52,53,54], eight in dyslipidemia [55,56,57,58,59,60,61,62], one in subjects with non-alcoholic fatty liver disease [63] and three in patients with ischemic coronary heart disease [64,65,66]. Trial intervention duration ranged from 3 weeks [46] to 24 months [50]. The sample size from all the trials ranged between 20 [29, 54, 67] and 160 participants [32, 33].

Table 1 Characteristics of the studies selected for meta-analysis of the effects of anthocyanins on cardiometabolic risk factors

A total of 23 trials compared provision of ACN supplement/extract in comparison with placebo or no intervention [23, 25, 28, 31,32,33, 43, 44, 46, 48,49,50,51,52, 54,55,56,57,58,59,60,61,62,63, 65].

Risk of bias in included studies

All trials provided data on losses to follow-up; eleven studies reported losses of > 10% [28, 30, 31, 34, 36, 40, 42, 44, 48, 55, 65] (Additional file 1: Figures S1 and S2). In thirteen trials, the information on the methods used for allocation concealment was unclear [23, 24, 30, 34,35,36,37,38, 42, 43, 47, 67, 68]. Three trials reported small sample size (< 15 subjects in each arm) [27, 67, 68].

Effect of anthocyanin on body mass index

Changes of BMI (kg/m2) after intervention with ACN were evaluated in twenty studies. From two studies that had multi-arm intervention, two pairs of arms were included in the analysis [30, 37]. Four studies included participants who were overweight or obese at baseline [27, 43,44,45].

The results of twenty trials showed that intervention with ACN resulted in a significant decrease in BMI (1,280 participants, mean difference (MD),− 0.21; 95% CI, − 0.38, − 0.04; P < 0.001) (Additional file 1: Figure S3).

The meta-regression analysis (random effects and maximum likelihood) with the difference of means as the dependent variable and the follow-up time (β = 0.01; 95% CI − 0.02, 0.042; p = 0.484), ACN dose (β = 0.00; 95% CI − 0.001, 0.00; p = 0.466) and BMI at baseline (β = 0.23; 95% CI − 0.04, 0.09; p = 0.490) as covariates showed no significant effects.

After conducting stratified analyses according to the type of ACN (supplement/extract vs. food), we found a statistically significant effect in both subgroups (Fig. 2). However, unlike supplement (MD: − 0.35; 95% CI − 0.47, − 0.23; P < 0.001), using ACN as powdered or food was accompanied by an increase in BMI in the intervention group as compared with control (MD: 0.1, 95% CI 0.04, 0.16; P = 0.001).

Fig. 2figure 2

Forest plot of subgroup analysis of the effect of anthocyanins compared with control on body mass index (kg/m2)

Sensitivity analyses were performed and made no remarkable change in overall results (MD: − 0.20, 95% CI − 0.39, − 0.04, P = 0.01). The results of leave-one-out analysis were shown in the (Additional file 1: Figure S4).

The Egger’s tests indicated no significant publication bias (P = 0.605).

Effect of anthocyanin on waist circumference

Eight trials compared the effect of ACN intake, as supplement/foods, versus control on waist circumference (WC). From one trial, two paired-arms were included in the analysis. Pooled data showed no statistically significant effect on WC (614 participants, MD: 0.78 cm, 95% CI − 0.43, 1.98 cm; P = 0.21) (Additional file 1: Figure S5).

Sensitivity analysis was performed for examining the effects of excluding trials with small sample sizes. The results were shown no significant effect (MD: − 0.09 cm, 95% CI − 1.25, 1.07; P = 0.88).

The results of leave-one-out analysis were shown in the (Additional file 1: Figure S6).

Effect of anthocyanin on body fat mass

Four trials compared the effect of ACN intervention with control on body fat mass (FM). Pooled data showed statistically significant effect on FM changes (150 participants, MD: − 0.3%, 95% CI − 0.42 to − 0.18%, p < 0.001) (Additional file 1: Figure S7).

The meta-regression analysis (random effects and maximum likelihood) with the difference of means as the dependent variable and the follow-up time (β = 0.01; 95% CI − 0.24, 0.27, p = 0.916) and used dose of ACN (β = 0.001; 95% CI − 0.01, 0.014, p = 0.853) as covariates showed the insignificant effects.

There were no data to allow for sensitivity analysis. The Egger’s tests indicated no significant publication bias (P = 0.881).

The results of leave-one-out analysis were shown in the (Additional file 1: Figure S8).

Effect of anthocyanin on fasting blood glucose

We pooled and analyzed data from 28 RCTs evaluating the effects of ACN intake on FBG. From three studies that had multi-arm intervention, two pairs of arms were included in the analysis [30, 37]. The forest plots for the effects of ACN intake on FBG are shown in the Additional file 1: Figure S9. The pooled MD was − 2.55 mg/dL (95%CI − 4.23, − 0.87), indicating a significantly greater reduction in FBG concentrations in the ACN intervention group than in the control (placebo) group.

The subgroup analyses were carried out to explore the impact of type of intervention (supplement/extract vs. food) and health status of participants (with diabetes vs. without diabetes). Significant reductions were observed in both subgroups of T2D patients (434 participants, MD: − 15.33 mg/dL; 95% CI − 25.8, − 7.86 mg/dL; P < 0.001) and participants without diabetes (1124 participants, MD: − 1.55 mg/dL; 95% CI − 2.38, − 0.72 mg/dL, P = 0.001). However, reduction of serum glucose concentrations was significantly more in the subjects with T2D (p < 0.001).

After a subgroup analysis, we found that FBG decreased significantly only in those who used ACN as supplement/extract (MD: − 3.13 mg/dL, 95% CI − 4.9, − 1.35 mg/dL; P = 0.004) (Fig. 3).

Fig. 3figure 3

Forest plot of subgroup

留言 (0)

沒有登入
gif