Urate-lowering effects of polyphenolic compounds in animal models: systematic review and meta-analysis

Protocol

This systematic review and meta-analysis will be conducted in strict accordance with the protocol registered with PROSPERO (CRD42024617024) and will adhere to the PRISMA guidelines (Moher et al., 2009).

Literature search strategy

The search covered databases from their inception until October 30, 2024. Databases searched include Embase, Wanfang database of Chinese academic institutions, PubMed, VIP database of Chinese technical journals, China National Knowledge Infrastructure (CNKI), and the Cochrane Library. The basic search framework involved terms for (“resveratrol”), (“chlorogenic acid”), (“ferulic acid”), (“punicalagin”), (“bergenin”), along with “ (serum uric acid)” and “(urinary uric acid)”. Details of the search strategy are provided in Supplementary Material 1.

Inclusion criteria

1. Study Type: in vivo animal experiments.

2. Subjects: animal models without restrictions on type, breed, gender, weight, or age.

3. Intervention: the experimental group received one of the polyphenols (RES, CGA, FA, PU, BER), while the control group received either a vehicle or a disease-inducing agent.

4. Outcomes: the primary outcome is serum uric acid, and the secondary outcome is urinary uric acid.

Exclusion criteria

1. Studies lacking customs-related data.

2. Studies with duplicate publications (retaining the study with the more comprehensive data set).

3. Research employing plant extracts containing the aforementioned five types of polyphenolic compounds or used in conjunction with other formulations.

4. Review articles.

5. In vitro experiments.

6. Conference abstracts.

7. Clinical studies.

8. Case reports.

Data extraction and management

Literature screening was independently conducted by two researchers, who systematically reviewed titles, abstracts, and full texts to identify studies meeting the predefined inclusion criteria. Discrepancies were resolved by consulting a third researcher. After preliminary screening, detailed examination of the full texts was undertaken based on the inclusion and exclusion criteria for further evaluation and data extraction. Data were independently extracted into a structured Excel database, capturing variables including author(s), year of publication, country of study, subject gender, species, age, body weight, dosage administered, model type, model induction pathway, details of experimental and control groups, intervention strategies, administration route, duration of intervention, and outcome measures. For all outcomes, data were extracted including sample sizes (N), means, standard deviations (SD), or standard errors of the mean (SEM). To standardize variability measures, the formula SD = SEM $\times \sqrt{}$N was employed to convert SEM values into SD. Outcome data were consistently recorded as means with their respective standard deviations. In instances where studies featured multiple experimental groups and a single control group, data from the control group were appropriately allocated across different comparisons. Each paired comparison was then included in the meta-analysis (Higgins et al., 2024). When study results were presented solely in graphical format, data extraction was performed using WebPlotDigitizer, and this extraction process was replicated three times to ensure precision. Efforts were made to contact authors for missing essential data; studies were excluded if authors did not respond, resulting in incomplete critical data.

Quality assessment of literature

The risk of bias within the included animal studies was evaluated by two researchers using the SYRCLE tool (Zeng et al., 2015). This assessment covered ten critical domains:

1. Sequence generation.

2. Baseline characteristics.

3. Allocation concealment.

4. Random housing.

5. Blinding of caregivers and investigators.

6. Random outcome assessment.

7. Blinding of outcome assessors.

8. Reporting of incomplete data.

9. Selective outcome reporting.

10. Other sources of bias.

Every study included was evaluated on an individual basis using these standards. A study was deemed “low risk” if it met all the criteria completely, reflecting high quality and a low probability of bias. Studies that met the criteria only partially were labeled as “unclear risk”, implying a moderate chance of bias. Those that did not meet the criteria at all were rated as “high risk”, signaling greater bias and reduced quality. Any conflicts were settled through consultation with a third researcher.

Statistical analysis

Statistical and meta-analytic computations were performed using R software version 4.4.0 and Review Manager (RevMan) 5.3. Outcomes were treated as continuous variables. When the studies shared uniform units of effect size and methods of measurement, the effect sizes were expressed using the 95% confidence intervals (CI) and weighted mean difference (WMD). In cases of non-uniformity, the standardized mean difference (SMD) and 95% CI were utilized instead. Heterogeneity among included studies was assessed using the I² statistic or the Q test. An I² value of 0%, 25%, 50%, and 75% indicated no, low, moderate, and high heterogeneity, respectively. For I² values ≥50%, sensitivity analyses were conducted by sequentially removing individual studies to observe changes in the I² value, exploring potential sources of heterogeneity, and assessing the stability of the results. If heterogeneity could not be addressed, a random effects model was employed along with subgroup analysis. Conversely, if the I² value was <50%, a fixed effects model was applied. For meta-analyses comprising ten or more studies, funnel plots and Egger’s test were conducted to assess publication bias, with a significance threshold set at p < 0.05 (Egger et al., 1997).

Included studies

Following the specified search strategy, a total of 1,060 articles were retrieved from six databases: PubMed (145 articles), Cochrane Library (11 articles), Embase (354 articles), CNKI (462 articles), VIP (20 articles), and Wanfang (68 articles). After the removal of duplicates, 266 articles were excluded. Screening of titles and abstracts led to the exclusion of 687 irrelevant publications, including reviews, reports, and clinical studies. Upon full-text review, 58 additional articles were excluded, leaving 49 articles for inclusion (Adikwu, Biradee & Ogungbaike, 2019; Alam, Sernia & Brown, 2013; Alhusaini et al., 2024; Anila et al., 2022; Bagul et al., 2018; Bagul et al., 2012; Bai et al., 2021; Che, 2024; Chen, 2013; Chowdhury et al., 2022; Da Silva et al., 2022; Erseçkin et al., 2022; Fan et al., 2009; Fang, 2015; Ferraz-Filha et al., 2017; Han et al., 2025; Hsu et al., 2020; Kaur et al., 2016; Lee et al., 2017; Li et al., 2021; Xu et al., 2019; Martins de Sá Müller et al., 2019; Meng et al., 2014; Palsamy & Subramanian, 2008; Peng et al., 2012; Qinglian, 2023; Rai, Bagul & Banerjee, 2020; Ren, 2015; Salau et al., 2023; Sammeturi et al., 2019; Sanjeev et al., 2019; Sarkar et al., 2019; Shi et al., 2012; Singh et al., 2021; Song, 2019; Sun Yadi et al., 2023; Wang, 2023; Xu et al., 2024; Xue et al., 2017; Yuan, 2017; Zhang et al., 2019; Zhang et al., 2022; Zhang et al., 2021; Zhang et al., 2017; Zhao et al., 2019; Zhao et al., 2016; Zhou & Chen, 2014; Zhou et al., 2024; Zhou et al., 2021). Two studies by Fan et al. (2009) and Alam, Sernia & Brown (2013) were further divided into four separate groups due to differences in modeling methodologies, animal species, and intervention durations, resulting in Fan(1)2009, Fan(2)2009, Alam(1)2013, and Alam(2)2013. A detailed flowchart of the literature screening process is depicted in Fig. 1.

Characteristics of included studies

The meta-analysis incorporated 49 studies conducted between 2008 and 2024, involving 81 independent control groups and 1,252 experimental animals. This cohort included 742 rats (517 Sprague-Dawley rats, 191 Wistar rats, 20 Albino rats, and 14 spontaneously hypertensive rats (SHR)) and 510 mice (318 Km mice, 69 Swiss mice, 107 C57BL/6J mice, and 16 ICR mice). Sprague-Dawley rats were utilized in 20 studies, Wistar rats in 10 studies, Albino rats in one study, and SHRs in one study. Km mice were used in nine studies, Swiss mice in five studies, C57BL/6J mice in four studies, and ICR mice in one study. The dosage range for polyphenolic compounds spanned from 5 to 1,000 mg/kg. Ten studies investigated three different dosages, another ten studies two dosages, and the remaining 29 studies focused on a single dosage. RES was used in 29 studies, FA in nine, CGA in eight, PU in two, and BER in one. Administration routes included oral gavage in 26 studies, oral administration in 18 studies, and intraperitoneal injection in five studies. The duration of intervention with polyphenolic compounds varied from 6 h to 20 weeks. Detailed characteristics of the included studies are presented in Table 1.

Quality assessment of studies

All studies were assessed as having unclear risk regarding baseline characteristics, allocation concealment, blinding of animal caregivers and researchers, and random outcome assessment. Only two studies provided detailed descriptions of random sequence generation, and 36 mentioned randomization in animal housing. One study noted that outcome assessors were blinded. Six studies reported incomplete data, as outcomes were not fully reported according to the experimental design. The risk of selective reporting and other potential biases was deemed low. Summaries and graphical representations of bias risks are shown in Figs. 2 and 3.

Effects of polyphenolic compounds on SUA

Impact of five polyphenolic compounds on SUA

A comprehensive review of fifty-one experiments, derived from forty-nine publications and involving 1,252 experimental animals, explored the effects of five distinct polyphenolic compounds on SUA levels in various animal models. The synthesized data indicated that these compounds generally lead to a reduction in SUA levels, albeit with significant variability among the results (SMD = −2.33, 95% CI [−2.73, −1.93], p < 0.0001, I² = 87%) (Fig. 4). A random effects model was utilized to manage the variability, and a subsequent sensitivity analysis corroborated the robustness of these findings (Fig. S1).

Further, subgroup analyses, stratified by animal species and specific disease models, demonstrated that the impact of polyphenols was markedly greater in Km mice (SMD = −5.56, p < 0.0001) and in models of liver disease (SMD = −5.56, p < 0.0001). This suggests a particularly strong efficacy of polyphenols in reducing SUA under these conditions. These subgroup findings also contributed to clarifying the sources of the noted heterogeneity (Table S1).

Impact of RES on SUA

An evaluation of thirty experiments, encompassing 732 experimental animals across twenty-nine publications, assessed the effects of RES on SUA levels in animal models. The pooled data from these studies reveal that RES significantly lowers SUA levels, although considerable heterogeneity was observed (SMD = −1.86, 95% CI [−2.28, −1.45], p < 0.0001, I² = 82%) (Fig. 5). A random effects model was again applied, and sensitivity analysis validated the consistency of these results (Fig. S2).

Detailed subgroup analysis highlighted that significant reductions in SUA levels were most prominent when the RES dosage ranged from 26 mg/kg to 50 mg/kg (SMD = −2.72, p < 0.0001), the treatment duration spanned from four to eight weeks (SMD = −2.69, p = 0.0002), the administration route was oral (SMD = −2.38, p = 0.0003), and when Swiss mice were used as the model (SMD = −3.29, p < 0.0001). Interestingly, the urate-lowering effect of RES displayed an inverted U-shaped nonlinear trend, with efficacy peaking within the 26–50 mg/kg dosage range and diminishing at higher doses. These observations not only underscore the potential of RES to optimize urate-lowering under specific conditions, but also illuminate the underlying factors contributing to the observed heterogeneity (Table S1).

Impact of CGA on SUA

Eight studies involving 174 experimental animals reported on the effects of CGA on SUA levels in animal models. Results indicate that CGA significantly reduces SUA levels, albeit with substantial heterogeneity (SMD = −2.31, 95% CI [−2.89, −1.73], p < 0.0001, I² = 57%) (Fig. 6). A random effects model was used, and sensitivity analysis confirmed the robustness of the results (Fig. S3).

Subgroup analysis showed significant reductions in SUA levels when CGA dosage was between 101 mg/kg and 300 mg/kg (SMD = −3.81, p < 0.0001), the duration was 2 to 4 weeks (SMD = −2.86, p < 0.0001), and in models using SD rats (SMD = −4.38). These subgroup analyses help to explain some of the observed heterogeneity (Table S1).

Impact of FA on SUA

Eleven experiments from nine publications involving 221 experimental animals reported on the impact of FA on SUA levels. The included studies showed that FA significantly lowers SUA levels, but with high heterogeneity (SMD = −2.82, 95% CI [−4.46, −1.19], p = 0.0007, I² = 93%) (Fig. 7). A random effects model was applied, and sensitivity analysis showed relatively stable results, although the exclusion of Sun2023 led to a notable change in the overall effect size, still within the original confidence interval (Fig. S4).

Subgroup analysis indicated significant reductions in SUA levels when FA dosage ranged from 101 mg/kg to 300 mg/kg (SMD = −9.97, p = 0.0001), the duration was 0 to 2 weeks (SMD = −18.93, p = 0.03), administered via oral gavage (SMD = −12.47, p < 0.0001), and in models using Km mice (SMD = −25.23, p < 0.0001). These subgroup analyses elucidate some sources of the observed heterogeneity (Table S1).

Impact of PU on SUA

Two studies involving 85 experimental animals reported the effects of PU on SUA levels in animal models. The pooled results demonstrate a significant reduction in SUA levels, albeit with high heterogeneity (SMD = −3.87, 95% CI [−5.99, −1.75], p = 0.0004, I² = 92%) (Fig. 8). A random effects model was employed, and sensitivity analysis confirmed the robustness of these results (Fig. S5).

Subgroup analyses indicated significant reductions in SUA levels with PU dosages between 101 mg/kg and 300 mg/kg (SMD = −7.93, p < 0.0001), durations of 0 to 2 weeks (SMD = −6.49, p < 0.0001), and in models using Km mice (SMD = −6.49, p < 0.0001). These subgroup analyses help to explain some of the observed heterogeneity (Table S1).

Impact of BER on SUA

A single study involving 40 experimental animals provided a descriptive analysis of the effects of BER on SUA levels in animal models. Compared to the control group, BER significantly reduced SUA levels (SMD = −8.51, 95% CI [−10.30, −6.73], p < 0.0001, I² = 0%) (Fig. 9).

The effect of polyphenolic compounds on UUA

Six studies involving 192 experimental animals reported the effects of polyphenolic compounds on UUA levels in animal models. The results indicate that these compounds significantly increase UUA levels, exhibiting high heterogeneity (SMD = 2.53, 95% CI [1.38–3.69], p < 0.01, I² = 91%) (Fig. 10). A random effects model was applied, and sensitivity analysis confirmed the robustness of the findings (Fig. S6).

Subgroup analyses showed significant increases in UUA levels when FA was used (SMD = 5.22, p = 0.004), in multisystem disease models (SMD = 6.35, p < 0.0001), and in Km mice (SMD = 2.90, p < 0.0001). These subgroup analyses partially account for the heterogeneity observed (Table S1).

Sensitivity analysis

Leave-one-out sensitivity analyses were conducted for all outcomes to observe whether the exclusion of individual studies significantly altered the effect sizes, thus assessing the stability of the results. Throughout these analyses, no significant changes in heterogeneity were observed across all measured outcomes in the meta-analyses (Figs. S1–S6).

Publication bias

To assess publication bias, funnel plots were generated for the overall effects of polyphenolic compounds on SUA and UUA, as well as for the specific effects of RES, CGA, FA, PU, and BER on SUA (Figs. S7–S13). Egger’s tests were conducted using R software, revealing significant publication bias for the overall impact of polyphenolic compounds on SUA and for the effects of RES on SUA, with p < 0.0001 for both.

Summary of results

A comprehensive synthesis of data from 49 studies reveals that the five selected polyphenolic compounds significantly reduce SUA levels and enhance UUA excretion in various animal models in comparison to control groups. These findings suggest that these specific polyphenols are effective in lowering uric acid levels, likely through mechanisms associated with increased renal excretion of uric acid. Moreover, subgroup analyses consistently indicate that these polyphenolic compounds decrease SUA levels across different disease models, suggesting that their uric acid-lowering effects are broadly applicable regardless of the underlying disease condition. However, the study exhibits significant heterogeneity. According to the subgroup analyses, variables such as the animal model used, dosage, duration of intervention, mode of administration, and disease model may contribute to this heterogeneity. Notably, subgroup analysis based on animal breeds revealed marked differences in the effects of some polyphenols, which could be attributed to genetic variations affecting uricase activity and uric acid metabolism among different breeds. Currently, experimental mouse models are primarily genetically induced or environmentally induced. Genetically induced models are limited by low survival rates and variable comparability due to complications, while environmentally induced models present challenges in direct comparison due to technical issues, such as variations in doses of inhibitors, induction periods, administration methods, and the measurement of serum uric acid concentrations. Such disparities may lead to inevitable heterogeneity in factors not related to the human condition (Lu et al., 2019). Additionally, individual meta-analyses were conducted to assess the specific effects of each polyphenolic compound (namely RES, CGA, FA, PU, and BER) on reducing SUA levels in animal models. The results confirm that each compound effectively reduces SUA levels.

In preclinical studies, the efficacy of RES in reducing SUA levels has been consistently affirmed. It was previously hypothesized that RES might reduce uric acid production through the inhibition of XO activity (Honda et al., 2014; Zhou et al., 2023). However, recent findings suggest that RES may also promote the enhanced excretion of UUA. Despite these insights, clinical trials examining the effects of RES continue to yield mixed results. For instance, studies conducted by Bo et al. (2016) and meta-analyses by Abdollahi et al. (2023) have observed no significant uric acid-lowering effects of RES in diabetic patients. Conversely, investigations by Zhou et al. (2023) have documented significant reductions in uric acid levels among individuals with dyslipidemia. This meta-analysis reveals that the urate-lowering effect of RES in animal models may be characterized by a potential “inverted U-shaped” nonlinear relationship. It appears that low to moderate doses of RES could reduce uric acid levels by dual inhibition of both XO and URAT1. The implications of high-dose RES are less clear; however, some studies indicate that such doses might lead to hepatotoxicity, increased oxidative stress, and suppression of natural killer (NK) cells, among other sub-toxic effects (Ferreira et al., 2024; Mendes-Braz et al., 2012; Petit et al., 2016). These adverse outcomes may contribute to the diminished uric acid-lowering effect observed at higher doses. Subgroup analysis within this meta-analysis demonstrates that RES shows pronounced efficacy in Swiss mice, particularly when administered orally in doses ranging from 26 mg/kg to 50 mg/kg over periods of 4 to 8 weeks. This suggests that an optimal dosing regimen for RES might involve oral administration at these doses, adjusted for body surface area (Huang Jihan et al., 2004), over a similar duration. However, the effectiveness of RES varies across biological species, and optimal conditions for its application in humans still require further clinical validation. Currently, the statistical power of dose grouping in these studies is inadequate, and uncertainties remain regarding the extrapolation of doses from animal models to humans. Consequently, the inverted U-shaped relationship identified should be considered a preliminary observation, necessitating additional mechanistic studies for verification.

CGA, primarily found in coffee, has been recognized early on for its beneficial effects in metabolic-related diseases (Clifford, 2000). Several clinical studies and animal experiments have reported its role in conditions such as obesity, hyperlipidemia, and diabetes. One animal study on mice with hyperuricemia found that CGA could protect against elevated levels of UA, BUN, Cr, and reduce inflammation through the inhibition of LPS, IL-1β, and TNF-α among other inflammatory factors and related signaling pathways (Zhou et al., 2021). Another animal study supported CGA’s ability to improve renal damage in hyperuricemic mice (Mehmood et al., 2020). Regarding the mechanisms by which CGA reduces uric acid, our results suggest it may be related to increased UUA excretion. Evidence from another animal study supports that CGA increases intestinal uric acid excretion by affecting the levels of intestinal urate transporters (ABCG2, GLUT9), combating hyperuricemia (Mehmood et al., 2019). An in vitro study showed that CGA lactone present in roasted coffee might inhibit XO activity (Honda et al., 2014). The mechanism of CGA’s uric acid reduction appears to be multi-pathway and multi-target. Subgroup results from this meta-analysis show that CGA’s uric acid-reducing effect is lower at 26 mg/kg to 50 mg/kg compared to 0 mg/kg to 25 mg/kg and 51 mg/kg to 100 mg/kg, which may be due to differences in animal models across groups.

In some studies, FA is considered to have a definitive uric acid-lowering effect, possibly including the inhibition of XO activity and increased expression of urate transport proteins (Lou et al., 2023; Zhang et al., 2022; Zhang et al., 2023). This aligns with our meta-analysis results; however, our subgroup analysis found that FA only showed a significant reduction in serum uric acid levels at doses of 101 mg/kg to 300 mg/kg, suggesting there may be a threshold effect for FA’s impact on SUA. The high heterogeneity (>50%) in FA dose subgroups could also mean that FA’s effects are influenced by other factors, necessitating further research for confirmation. Some studies have found that FA could cause renal damage in conditions like diabetic nephropathy, renal cancer, and chronic kidney disease in rats (Bagul et al., 2012; Cai, 2010), therefore, we do not currently recommend the widespread use of FA in uric acid-lowering drug development.

Current research on PU is relatively limited, but some studies have mentioned that PU exhibits a serum uric acid-lowering effect by upregulating some uric acid-excreting transporters and downregulating some that absorb uric acid (Masuoka, 2021). Our results indicate that PU significantly reduces serum uric acid. However, in the 51 mg/kg to 100 mg/kg subgroup, the effect of PU on reducing serum uric acid was unclear, and this group exhibited high heterogeneity (>50%). An in-depth analysis of the two studies in this subgroup suggests that the variance in results could be due to differences in disease models, animal models, intervention durations, individual differences among animals, and publication bias.

Given the limited number of studies on BER, we only included one study, which does not significantly inform the analysis of BER’s impact on SUA. BER is primarily found in a few plants like rockfoil, which may be one reason for the scarcity of research on this compound.

Strengths and limitations

This research represents the inaugural systematic review and meta-analysis dedicated to examining the influence of polyphenolic compounds on uric acid levels within animal models. It encompassed an extensive literature search and meticulous analysis. The implementation of subgroup and sensitivity analyses illuminated potential sources of heterogeneity and evaluated the influence of individual studies on the aggregate outcomes, thus providing invaluable reference points for subsequent inquiries.

Nevertheless, the quality of the studies incorporated was predominantly subpar, characterized by ambiguous risks linked to the randomization processes, allocation concealment, baseline characteristics, and the blinding of researchers. Such deficiencies likely contribute to the observed heterogeneity of research findings. While these limitations could potentially exaggerate effect sizes, the consistency of results across varied models and compounds lends credibility to their biological feasibility. We recommend exercising caution in the interpretation of these findings and advocate for enhanced reporting standards in preclinical research. Both the Egger test and funnel plots revealed indications of publication bias, which may compromise the reliability of the uric acid-lowering effects attributed to polyphenolic compounds and could potentially skew the outcomes of this meta-analysis. This optimistic estimation of effect size necessitates validation through extensive, large-scale preclinical trials in the future. During the data extraction phase, the prevalence of studies presenting data in graphical formats led to unavoidable inaccuracies in manual data extraction. Additionally, the small sample sizes prevalent in the included studies undermine the reliability of both the meta-analysis and subgroup analysis results. Although the duration of interventions accounts for some heterogeneity, other potential variables—such as undisclosed rearing conditions, variations in genetic backgrounds of different strains, and disparities in the isolated compounds and their natural sources—remain to be thoroughly investigated through standardized reporting in future research endeavors. The subgroup analyses conducted in this study were constrained by the reporting dimensions of the original data. We advocate for future animal studies to adhere to the STAIR guidelines for standardized reporting of intervention details to facilitate a more precise attribution of heterogeneity (Savitz & Fisher, 2007). Additionally, the lack of specificity regarding the type of disease in animal models introduced further heterogeneity and led to inquiries into whether the application of the five polyphenols was influenced by disease type to some extent. The analysis did not segregate rodent models into distinct groups, which heightened heterogeneity and posed significant challenges in discussing factors influencing the effects of polyphenols; however, this did not appear to markedly affect the overall or subgroup-specific outcomes.