ABSTRACT
Chronic low‐grade inflammation, adipokine dysregulation, and elevated liver enzymes are associated with metabolic disease. Given their rich content of polyphenols, dietary fiber, and other bioactive compounds, apples and apple‐derived products may modulate these biomarkers, although trial findings remain inconsistent. An advanced database search was conducted up to September 1, 2026, to identify eligible randomized controlled trials (RCTs). Outcomes of interest included C‐reactive protein (CRP), interleukin‐6 (IL‐6), tumor necrosis factor‐alpha (TNF‐α), adiponectin, alanine aminotransferase (ALT), aspartate aminotransferase (AST), and alkaline phosphatase (ALP). Pooled effect sizes were calculated as weighted mean differences (WMDs) with 95% confidence intervals (CIs) using a random‐effects model. A total of 19 RCTs were included in the meta‐analysis. The pooled analysis showed that apple and apple‐derived products did not significantly affect CRP (WMD: −0.03 mg/L; p = 0.388), IL‐6 (WMD: −0.14 pg/mL; 95% CI: −0.31, 0.02; p = 0.082), TNF‐α (WMD: −0.09 pg/mL; p = 0.432), adiponectin (WMD: 0.22 mg/dL; p = 0.450), ALT (WMD: −0.92 IU/L; p = 0.234), ALP (WMD: −2.81 IU/L; p = 0.089). However, a significant reduction was observed for AST (WMD: −1.56 IU/L; p = 0.006). Exploratory subgroup findings showed that some benefits were observed more frequently in specific contexts, including particular product forms (e.g., extracts or vinegar), longer intervention duration, and specific participant subgroups. Apple and apple‐derived products did not significantly affect most inflammatory biomarkers, adiponectin, or liver enzymes, except for a significant reduction in AST. Given the low to very low certainty of evidence for most outcomes, further well‐designed RCTs are needed to clarify their potential effects.
Keywords: adiponectin, apple, inflammation, liver enzymes, Malus domestica , meta‐analysis, review
Summary of the systematic review and meta‐analysis of 19 randomized controlled trials evaluating the effects of apple and apple‐derived products on inflammatory biomarkers, adiponectin, and liver enzymes in adults. Apple interventions significantly reduced AST, whereas no statistically significant effects were observed for CRP, IL‐6, TNF‐α, adiponectin, ALT, or ALP.

Abbreviations
- ALP
alkaline phosphatase
- ALT
alanine aminotransferase
- AST
aspartate aminotransferase
- CI
confidence interval
- CRP
C‐reactive protein
- IL‐6
interleukin‐6
- RCT
randomized controlled trial
- TNF‐α
tumor necrosis factor‐alpha
- WMD
weighted mean difference
1. Introduction
Chronic low‐grade inflammation contributes to the development and progression of numerous non‐communicable and metabolic diseases (Cifuentes et al. 2024). Circulating inflammatory biomarkers, such as C‐reactive protein (CRP), interleukin‐6 (IL‐6), and tumor necrosis factor‐alpha (TNF‐α), are established indicators of systemic inflammation and are closely associated with cardiometabolic risk (Ruan et al. 2024). In addition, adipokines (Kirichenko et al. 2022; Clemente‐Suárez et al. 2023) and liver enzymes, including alanine aminotransferase (ALT), aspartate aminotransferase (AST), and alkaline phosphatase (ALP), are important biomarkers of metabolic homeostasis and liver function, with abnormal levels linked to obesity, insulin resistance, and chronic liver disease (Clemente‐Suárez et al. 2023; Liu et al. 2021). Consequently, identifying safe, effective, and readily accessible dietary interventions that can modulate these pathophysiological pathways has become a major focus of clinical nutrition and preventive medicine. Growing evidence indicates that dietary interventions and bioactive food‐derived products can improve inflammatory and cardiometabolic health, highlighting their potential as complementary strategies for preventing and managing metabolic disorders (Setayesh et al. 2025; Karimi, Javadi, Sharifi, et al. 2025; Karimi, Javadi, Hamzavi, et al. 2025; Karimi et al. 2026).
Apple ( Malus domestica L.) is one of the most frequently consumed fruits globally, and it is recognized not only for its availability and sensory appeal but also as a significant source of bioactive compounds in the human diet (Garmo et al. 2026; Wang et al. 2014). The health‐promoting effects of apples are mostly attributed to a diverse profile of polyphenols, such as flavan‐3‐ols, hydroxycinnamic acids, dihydrochalcones, flavonols, and anthocyanins, as well as soluble fiber (pectin) (Bars‐Cortina et al. 2020; Lee et al. 2003).
Findings of experimental studies indicate that these apple‐derived polyphenols are believed to enhance bioactivity, exerting pleiotropic effects such as the inhibition of reactive oxygen species (ROS) production, the modulation of cell signaling cascades involved in inflammatory cytokine secretion, and the improvement of endothelial function by mediated pathways of nitric oxide (Bondonno et al. 2018; Del Bo et al. 2015). In addition, specific components such as pectin can enhance the gut microbiota, leading to improved lipid metabolism, glycemic control, and systemic inflammation through complex gut‐brain signaling (Koutsos et al. 2017; Koutsos et al. 2015).
While epidemiological studies consistently report an inverse association between apple intake and the risk of chronic metabolic and cardiovascular diseases (Hodgson et al. 2016; Sandoval‐Ramírez et al. 2020), findings from human clinical trials remain inconsistent (Auclair et al. 2010; Ravn‐Haren et al. 2013). Discrepancies in the literature may stem from heterogeneity in the types of apple products, including whole fruit, skin, and pomace, as well as apple juice and purified extracts, as well as variations in dosage, study duration, and the metabolic populations investigated (Koutsos et al. 2020; Pedret et al. 2024). Although individual trials have explored the impact of apple consumption on specific outcomes, there is a lack of comprehensive evidence synthesizing these effects on the triad of serum inflammatory biomarkers, adiponectin, and liver enzymes (Hollands et al. 2018; Joseph et al. 2016).
A recent systematic review by Del Bo' et al. (2026) reviewed 38 human intervention studies and found that apples and apple‐derived products may improve several cardiometabolic and functional markers, although substantial heterogeneity limited definitive conclusions. However, their review did not quantitatively synthesize evidence specifically for inflammatory biomarkers, adiponectin, and liver enzymes.
Given apples' potential as an affordable, accessible dietary strategy, a comprehensive evaluation of their clinical effects is needed to better inform evidence‐based nutritional recommendations. Therefore, this systematic review and meta‐analysis of randomized controlled trials (RCTs) aimed to assess the effects of apples and/or apple‐derived products on serum inflammatory biomarkers, adiponectin, and liver enzymes in adults. Additionally, this study explored potential sources of variability across studies and provided insights into factors that may influence the effectiveness of these interventions.
2. Methods
2.1. Study Protocol and Registration
This systematic review and meta‐analysis was conducted in accordance with the Cochrane Handbook for Systematic Reviews of Interventions (Higgins et al. 2019) and reported following the Preferred Reporting Items for Systematic Reviews and Meta‐Analyses (PRISMA) guidelines (Page et al. 2021). The review protocol was prospectively registered in the International Prospective Register of Systematic Reviews (PROSPERO Registration No. CRD420261463680). The research question and eligibility criteria were developed using the PICOS (Population, Intervention, Comparison, Outcomes, and Study Design) framework (Amir‐Behghadami and Janati 2020). The PICOS framework and detailed inclusion and exclusion criteria used to determine study eligibility are presented in Table 1.
TABLE 1.
PICOS framework and inclusion/exclusion criteria.
| PICOS domain | Inclusion criteria | Exclusion criteria |
|---|---|---|
| Population (P) | Adults aged ≥ 18 years, regardless of sex, geographical region, baseline health status, or comorbidities | Participants aged < 18 years |
| Intervention (I) | Apple or apple‐derived products, including whole apples, apple juice, apple vinegar, apple extracts, or other apple‐based preparations, administered as the sole intervention | Interventions in which apple products were combined with other active ingredients and their independent effects could not be isolated |
| Comparison (C) | Placebo, control group, no supplementation, or no‐intervention group | Uncontrolled or non‐randomized intervention studies |
| Outcomes (O) | At least one predefined outcome: CRP, IL‐6, adiponectin, ALT, AST, or ALP | Studies reporting none of the predefined outcomes |
| Study design (S) | RCT with parallel‐group or crossover designs | Observational studies (cohort, case–control, cross‐sectional, or ecological), case reports, reviews, editorials, letters, conference abstracts, non‐randomized studies, and uncontrolled intervention studies |
| Quantitative data | Sufficient numerical data to calculate effect sizes | Insufficient data for quantitative synthesis after attempts to obtain additional information |
| Publication and study availability | Full‐text studies meeting all other eligibility criteria; no restrictions on publication year or language | Duplicate publications (only the most complete report retained), inaccessible full texts, and gray literature (e.g., dissertations, theses, conference proceedings, and unpublished studies), which was not searched |
Abbreviations: ALP, alkaline phosphatase; ALT, alanine aminotransferase; AST, aspartate aminotransferase; CRP, C‐reactive protein; IL‐6, interleukin‐6; RCT, randomized controlled trial.
2.2. Eligibility Criteria
Studies were considered eligible if they met all of the following inclusion criteria: (1) RCTs (parallel or crossover design); (2) conducted in adults aged ≥ 18 years; (3) evaluated the effects of apple or apple‐derived products (e.g., whole apples, apple juice, apple vinegar, apple extracts, or other apple‐based preparations) administered as the sole intervention; (4) included a placebo, control, or no‐treatment group for comparison; (5) reported data for at least one of the predefined outcomes; and (6) provided sufficient quantitative data to calculate effect sizes. No restrictions were imposed regarding publication year, language, sex, geographical region, baseline health status, or comorbidities.
Studies were excluded if they met any of the following criteria: (1) observational studies (cohort, case–control, cross‐sectional, or ecological studies), case reports, conference abstracts, review articles, editorials, letters, or animal and in vitro studies; (2) non‐randomized or uncontrolled intervention studies; (3) studies involving participants < 18 years of age; (4) interventions in which the independent effect of apple products could not be isolated because they were combined with other active ingredients; (5) studies that did not report any of the predefined outcomes or lacked sufficient data for quantitative synthesis after attempts to obtain additional information; (6) duplicate publications, for which only the most complete report was retained. Gray literature (e.g., dissertations, theses, conference proceedings, and unpublished studies) was not searched, and studies with inaccessible full texts were therefore not considered for inclusion.
2.3. Search Strategy
A comprehensive and systematic literature search was conducted to identify all relevant RCTs. Electronic databases, including PubMed/MEDLINE, Scopus, and Web of Science, were searched from the inception of each database through September 1st 2026, without restrictions on publication year or language. The search strategy incorporated keywords and synonyms related to apple, inflammatory biomarkers, adiponectin, liver enzymes, and RCT. Boolean operators (AND/OR), truncation symbols, and database‐specific search filters were applied as appropriate to maximize sensitivity. The detailed search keywords and search strategies for all three databases are provided separately in Supplementary Tables 1 and 2. To ensure comprehensive study identification, Google Scholar and the reference lists of all eligible articles and relevant review papers were manually screened for additional studies that may have been missed during the electronic database search.
2.4. Data Extraction
Two reviewers (S.H. and M.M.K.) independently extracted data from all eligible studies using a standardized data extraction form. Any discrepancies between reviewers were resolved through discussion and, when necessary, consultation with a third reviewer (O.A.) until consensus was reached. The following information was extracted from each included trial: first author, publication year, country, study design, sample size, participant characteristics (including mean age, sex distribution, body mass index [BMI], and health status), details of the intervention (type of apple or apple‐derived product, dosage, and duration), characteristics of the control group, and outcome data. For each outcome of interest, baseline and post‐intervention means, mean changes (when available), corresponding measures of variability (standard deviation [SD], standard error [SE], or 95% confidence interval [CI]), and the number of participants analyzed were extracted for serum inflammatory biomarkers (CRP, IL‐6, and TNF‐α), adiponectin, and liver enzymes (ALT, AST, and ALP). When outcome data were incomplete or reported in an unsuitable format, appropriate statistical conversions were performed according to the Cochrane Handbook for Systematic Reviews of Interventions (Higgins et al. 2019) to obtain comparable effect estimates for the meta‐analysis.
2.5. Statistical Analysis
All statistical analyses were performed using Stata software, version 19 (Stata Corp., College Station, TX), with statistical significance defined as a two‐tailed p‐value < 0.05. To synthesize the findings, a random‐effects model was employed to calculate pooled weighted mean differences (WMDs) with corresponding 95% confidence intervals (CIs), thereby accounting for potential between‐study heterogeneity (DerSimonian and Laird 1986). Mean difference outcomes were determined by comparing changes from baseline to post‐intervention between the intervention and comparison groups. For studies omitting the standard deviation (SD) of the mean change, SDs were calculated using the following Cochrane‐approved formula (Borenstein et al. 2009):
When the SD of the change score was unavailable, it was calculated using a standard formula and a correlation coefficient (r) of 0.8. SEs, 95% CIs, and IQRs were converted to SDs using the approach described by Hozo et al. (Hozo et al. 2005). Meta‐analyses used a random‐effects model to account for between‐study heterogeneity. Heterogeneity was assessed with Cochran's Q test and quantified using the I 2 statistic. In accordance with Cochrane guidance (Ioannidis 2008), I 2 values of 0%–40%, 40%–60%, 60%–80%, and 80%–100% were interpreted as indicating low, moderate, substantial, and considerable heterogeneity, respectively (Ioannidis 2008).
For crossover RCTs, data from the intervention and comparison periods were treated as paired observations when sufficient within‐participant variance information was available; where paired variance estimates were not reported, the available variance data were used according to the information provided in the original study.
To explore potential sources of heterogeneity, pre‐specified exploratory subgroup analyses were conducted based on participant characteristics including geographical region, sex, baseline BMI categories (kg/m2), trial duration (< 8 weeks vs. ≥ 8 weeks), and intervention type (apple extract vs. apple vinegar).
Furthermore, random‐effects meta‐regression and non‐linear dose–response regression analyses were conducted to evaluate potential moderating effects of intervention study duration on the specified outcomes, thereby facilitating the synthesis of correlated dose–response data across studies (Xu and Doi 2018; Xie et al. 2021). Potential publication bias was rigorously assessed using visual inspection of funnel plots, Egger's linear regression, and Begg's rank correlation tests, with the trim‐and‐fill method applied to address any detected asymmetry (Shi and Lin 2019). Finally, the robustness of the pooled estimates was evaluated through sensitivity analyses using the leave‐one‐out method (Begg and Mazumdar 1994; Egger et al. 1997).
2.6. Risk of Bias Assessment
Risk of bias in the included papers was independently evaluated by two authors (S.H. and M.M.K.) using the RoB‐2 tool (the revised Cochrane risk‐of‐bias tool for randomized trials). RoB‐2 assesses bias in five distinct domains: the randomization process, deviations from intended interventions, missing outcome data, measurement of the outcome, and selection of the reported result. The judgments within each domain lead to a general risk of bias: “low risk of bias,” “some concerns,” or “high risk of bias” (Sterne et al. 2019).
2.7. Certainty of Evidence Assessment
The overall certainty of evidence across RCTs was evaluated separately by two reviewers (O.A. and M.K.) utilizing the GRADE guidelines (Guyatt et al. 2008). As a result, the quality of evidence was classified into four categories: high, moderate, low, and very low. These criteria ensured a standardized and transparent assessment of the strength of the evidence.
3. Results
3.1. Study Selection
PRISMA Flow chart of the study selection process for included studies in the systematic review is illustrated in Figure 1. A total of 770 records were identified through database searching, including PubMed (n = 125), Web of Science (n = 257), and Scopus (n = 388). After removing 191 duplicate records, 579 articles remained for title and abstract screening. Of these, 555 records were excluded because they were review articles, non‐human studies, or had irrelevant titles or abstracts. The full texts of 24 potentially eligible studies were assessed, and five were subsequently excluded due to non‐randomized design, insufficient outcome data, or lack of an appropriate control group. Ultimately, 19 studies (Auclair et al. 2010; Ravn‐Haren et al. 2013; Koutsos et al. 2020; Pedret et al. 2024; Akazome et al. 2010; Alzoufairi et al. 2025; Barnett et al. 2021; Barth et al. 2012; Chai et al. 2012; Cicero et al. 2017; Gasper et al. 2014; Gheflati et al. 2019; Guo et al. 2008; Liddle et al. 2021; Nagasako‐Akazome et al. 2007; Panetta et al. 2013; Pushpass et al. 2023; Saarenhovi et al. 2017; Tenore et al. 2018) met the eligibility criteria and were included in the systematic review and meta‐analysis.
FIGURE 1.

PRISMA Flow chart of the study selection process for the inclusion of randomized controlled trials (RCTs) in the systematic review and meta‐analysis of the effects of apples and apple‐derived products on serum inflammatory biomarkers, adiponectin, and liver enzymes in adults.
3.2. Study Characteristics
A total of 19 RCTs (Auclair et al. 2010; Ravn‐Haren et al. 2013; Koutsos et al. 2020; Pedret et al. 2024; Akazome et al. 2010; Alzoufairi et al. 2025; Barnett et al. 2021; Barth et al. 2012; Chai et al. 2012; Cicero et al. 2017; Gasper et al. 2014; Gheflati et al. 2019; Guo et al. 2008; Liddle et al. 2021; Nagasako‐Akazome et al. 2007; Panetta et al. 2013; Pushpass et al. 2023; Saarenhovi et al. 2017; Tenore et al. 2018) comprising 24 intervention arms were included in this systematic review (Table 2). The studies were published between 2007 and 2025 and were conducted across diverse geographical regions, including Japan, China, France, Germany, Denmark, the USA, the UK, Finland, Italy, Iran, Canada, New Zealand, Ireland, and Spain. Most trials employed a parallel‐group design, while the remainder used a crossover design, with the majority being placebo‐controlled and several incorporating double‐ or single‐blind methodologies. The studies enrolled a broad range of participants, including healthy adults, overweight or obese individuals, patients with hypercholesterolemia, type 2 diabetes mellitus with dyslipidemia, borderline or mild hypertension, and non‐diabetic obese participants. Sample sizes ranged from 29 to 100 participants, with intervention groups ranging from 13 to 57 participants, and the mean age of participants ranged from approximately 36 to 64 years. Baseline BMI ranged from 20.8 to 33.3 kg/m2, representing normal‐weight, overweight, and obese populations. Intervention durations ranged from 10 days to 52 weeks, although most studies lasted between 4 and 12 weeks. A wide variety of apple‐derived interventions were evaluated, including whole apples, dried apples, lyophilized apples, apple juice (clear and cloudy), apple puree, apple pomace, apple cider vinegar/apple vinegar, and apple polyphenol extracts (including Annurca apple and Malus domestica extracts), with daily doses varying according to the intervention type. Control groups received placebo preparations, no treatment, alternative apple varieties, sugar‐matched or isocaloric control beverages, dried plums, pomegranate juice, aspirin, or cornflakes with milk plus placebo capsules, depending on the study design.
TABLE 2.
Characteristics of the 19 included randomized controlled trials (RCTs) in the systematic review and meta‐analysis of the effects of apples and apple‐derived products on serum inflammatory biomarkers, adiponectin, and liver enzymes in adults.
| Studies | Country | Study design | Health status of participant | Sex | Sample size | Trial duration (week) | Age (year) | BMI (kg/m2) | Intervention | ||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| IG | CG | IG | CG | IG | CG | IG | CG | ||||||
| Nagasako‐Akazome et al. (2007) | Japan | RCT, Parallel, DB, PC | Healthy and moderately obese | M, F | 24 | 24 | 12 | 47.8 ± 10.8 | 46.5 ± 10.8 | 25.9 ± 1.9 | 25.9 ± 2.2 | Apple polyphenol extract (600 mg/day) | Placebo |
| Guo et al. (2008) | China | RCT, Parallel | Elderly | M: 20, F: 6 | 13 | 13 | 4 | 62.8 ± 4.7 | 64.1 ± 4.3 | 24.2 ± 2.5 | 23.0 ± 2.5 | Apple juice (250 mL/day) | Pomegranate |
| Auclair et al. (2010) | France | RCT, Crossover, DB | Hyper‐cholesterolemia | M: 30 | 30 | 30 | 4 | 52.6 ± 5.5 | 52.6 ± 5.5 | 25.7 ± 2.6 | 25.7 ± 2.6 | Lyophilized apples (Marie Me'nard apple) (40 g/day) | Golden Delicious apple |
| Akazome et al. (2010) | Japan | RCT, Parallel, DB, PC | Healthy | M: 56, F: 38 | 47 | 47 | 12 | 45.3 ± 8.6 | 44.9 ± 9.1 | 27.1 ± 1.5 | 27.0 ± 1.4 | Apple polyphenol extract (340 g/bottle (containing 600 mg apple polyphenols) three times/day) | Placebo |
| Akazome et al. (2010) | Japan | RCT, Parallel, DB, PC | Healthy | M: 20, F: 10 | 15 | 15 | 4 | 40.4 ± 10.4 | 41.6 ± 9.8 | 23.3 ± 3.0 | 23.4 ± 2.9 | Apple polyphenol extract (340 g/bottle (containing 600 mg apple polyphenols) three times/day) | Placebo |
| Chai et al. (2012) | USA | RCT, Parallel | Healthy postmenopausal women | F:160 | 45 | 55 | 52 | 55.6 ± 5.0 | 57.5 ± 4.01 | 24.8 ± 4.1 | 24.9 ± 4.6 | Dried apple (75 g/day) | 100 g dried plum |
| Barth et al. (2012) | Germany | RCT, Parallel, SB | Non‐diabetic obese men | M: 68 | 38 | 30 | 4 | 49 (23–69) | 49 (23–69) | 31.1 ± 3.6 | 30.5 ± 3.0 | Cloudy apple juice (750 mL/day) | 750 mL of control beverage (isocaloric)/day |
| Ravn‐Haren et al. (2013) | Denmark | RCT, Crossover, SB | Healthy | M: 9, F: 14 | 23 | 23 | 20 | 36.2 ± 17.9 | 36.2 ± 17.9 | 22.3 ± 2.59 | 22.3 ± 2.59 | Whole fresh apples (550 g/day) | No treatment |
| Ravn‐Haren et al. (2013) | Denmark | RCT, Crossover, SB | Healthy | M: 9, F: 14 | 23 | 23 | 20 | 36.2 ± 17.9 | 36.2 ± 17.9 | 22.3 ± 2.59 | 22.3 ± 2.59 | Apple pomace (22 mg/day) | No treatment |
| Ravn‐Haren et al. (2013) | Denmark | RCT, Crossover, SB | Healthy | M: 9, F: 14 | 23 | 23 | 20 | 36.2 ± 17.9 | 36.2 ± 17.9 | 22.3 ± 2.59 | 22.3 ± 2.59 | Cloudy apple juice (500 mL/day) | No treatment |
| Ravn‐Haren et al. (2013) | Denmark | RCT, Crossover, SB | Healthy | M: 9, F: 14 | 23 | 23 | 20 | 36.2 ± 17.9 | 36.2 ± 17.9 | 22.3 ± 2.59 | 22.3 ± 2.59 | Clear apple juice (500 mL/day) | No treatment |
| Panetta et al. (2013) | USA | RCT, Parallel, DB, PC | Non‐diabetics | M: 63, F: 34 | 50 | 47 | 8 | 57.7 ± 9.33 | 56.1 ± 12.58 | 28.3 ± 5.87 | 26.8 ± 4.19 | Apple cider vinegar (30 mL/day) | Placebo |
| Gasper et al. (2014) | UK | RCT, Crossover | Healthy | M: 13, F: 12 | 25 | 25 | 4 | 42 ± 11 | 42 ± 11 | 25.9 ± 4.2 | 25.9 ± 4.2 | Apple puree (230 g/day [containing 100 mg epicatechin]) | Aspirin (75 mg/day) |
| Gasper et al. (2014) | UK | RCT, Crossover | Healthy | M: 13, F: 12 | 26 | 25 | 4 | 42 ± 11 | 42 ± 11 | 25.9 ± 4.2 | 25.9 ± 4.2 | Apple puree (230 g/day [containing 25 mg epicatechin]) | Aspirin (75 mg/day) |
| Saarenhovi et al. (2017) | Finland | RCT, Crossover, DB, PC | Borderline hypertension or unmedicated mild hypertension | M: 26, F: 34 | 57 | 57 | 4 | 55.6 ± 7.9 | 54.9 ± 6.4 | 24.7 ± 3.0 | 26.3 ± 3.5 | Apple polyphenol extract (330 mg/day (containing 100 mg of epicatechin)) | Placebo |
| Cicero et al. (2017) | Italy | RCT, parallel, DB, PC | Overweight (BMI > 25) | M, F | 31 | 31 | 8 | 63.3 ± 7.5 | 63.2 ± 11.6 | 25–30 | 25–30 | Apple polyphenol extract ( Malus Domestica ) (300 mg/day) | Placebo |
| Tenore et al. (2018) | Italy | RCT, Crossover, TB, PC | Healthy | M: 30, F: 20 | 50 | 50 | 10 days | 43.5 ± 10.2 | 43.5 ± 10.2 | 20.8 ± 3.1 | 20.8 ± 3.1 | Annurca apple polyphenolic extract (1000 mg/day) | Placebo |
| Gheflati et al. (2019) | Iran | Parallel, C | T2DM & dyslipidemia | M: 20, F: 42 | 32 | 30 | 8 | 49.47 ± 8.02 | 52.1 ± 7.87 | 28.96 ± 4.31 | 28.85 ± 4.24 | Apple vinegar (20 mL/day) | No treatment |
| Koutsos et al. (2020) | UK | RCT, Crossover | Healthy with mild hyper‐cholesterolemia | M: 17, F: 23 | 40 | 40 | 8 | 51 ± 11 | 51 ± 11 | 25.3 ± 3.7 | 25.3 ± 3.7 | Whole apple (2 apples [340 g]/day) | 500 mL of sugar‐matched apple control beverage |
| Liddle et al. (2021) | Canada | RCT, Parallel, PC | Obese and overweight | M: 14, F: 30 | 22 | 22 | 6 | 48 ± 14.54 | 42.9 ± 15.0 | 33.2 ± 6.09 | 33.3 ± 5.62 | Whole Gala apple (600 g [3 whole Gala apples (∼200 g edible parts)/day]) | Placebo |
| Barnett et al. (2021) | New Zealand | RCT, Crossover | Healthy | M: 5, F: 20 | 25 | 25 | 2 | F: 39.8, M: 47.3 | F: 38, M: 39.7 | F: 23.3, M: 22.8 | F: 23.4, M: 27.4 | Red‐fleshed apple (1 portion (equivalent to 1 apple) daily) | White‐fleshed apple |
| Pushpass et al. (2023) | Ireland | RCT, Parallel, DB, C | Healthy | M: 15, F: 46 | 16 | 15 | 8 | 55 ± 16 | 49 ± 15.49 | 25.7 ± 3.87 | 24.9 ± 3.65 | Renetta apple (2 apples/day) | 40 g Cornflakes with 180 mL semi‐skimmed milk and 2 placebo capsules |
| Pedret et al. (2024) | Spain | RCT, Parallel, C | Hyper‐cholesterolemia | M: 51, F: 70 | 40 | 41 | 6 | 46.7 ± 16.3 | 49.8 ± 13.6 | 26.3 ± 3.8 | 24.6 ± 3.2 | Red‐fleshed apples (80 g/day) | White‐fleshed apples |
| Alzoufairi et al. (2025) | UK | RCT, Parallel, PC | Healthy | M: 14, F: 45 | 15 | 14 | 8 | 54 ± 3 | 50 ± 3 | 25.5 ± 0.9 | 24.7 ± 0.9 | Renetta Canada apples (2 Renetta Canada apples/day with 2 placebo capsules) | Placebo |
Abbreviations: BMI, body mass index; CG, control group; DB, double blind; F, female; IG, intervention group; M, male; NR, not reported; PC, placebo‐controlled; RCT, randomized controlled trial; T2DM, type 2 diabetes mellitus.
3.3. Risk of Bias Assessment
The risk of bias assessment of the 19 included RCTs (Auclair et al. 2010; Ravn‐Haren et al. 2013; Koutsos et al. 2020; Pedret et al. 2024; Akazome et al. 2010; Alzoufairi et al. 2025; Barnett et al. 2021; Barth et al. 2012; Chai et al. 2012; Cicero et al. 2017; Gasper et al. 2014; Gheflati et al. 2019; Guo et al. 2008; Liddle et al. 2021; Nagasako‐Akazome et al. 2007; Panetta et al. 2013; Pushpass et al. 2023; Saarenhovi et al. 2017; Tenore et al. 2018) is presented in Table 3. Overall, 10 RCTs were judged to have a high risk of bias, 8 RCTs were rated as having some concerns, and only 1 RCT was considered to have a low risk of bias. The most frequent sources of bias were related to the randomization process (Domain 1), deviations from intended interventions (Domain 2), and missing outcome data (Domain 3). In contrast, the majority of studies were judged to be at low risk of bias for the measurement of outcomes (Domain 4) and the selection of the reported results (Domain 5). Overall, these findings indicate that methodological limitations were common among the included trials and should be considered when interpreting the pooled estimates.
TABLE 3.
Risk of bias (RoB) assessment of the included 19 randomized controlled trials (RCTs) in the systematic review and meta‐analysis of the effects of apples and apple‐derived products on serum inflammatory biomarkers, adiponectin, and liver enzymes in adults.
| Studies | D1 | D2 | D3 | D4 | D5 | Overall RoB |
|---|---|---|---|---|---|---|
| Nagasako‐Akazome et al. (2007) | High | Low | Low | Low | High | High |
| Guo et al. (2008) | High | High | Low | Low | Low | High |
| Auclair et al. (2010) | High | Low | Low | Low | Low | Some concerns |
| Akazome et al. (2010) | High | Low | Low | Low | Low | Some concerns |
| Chai et al. (2012) | Low | High | High | Low | Low | High |
| Barth et al. (2012) | High | Low | High | Low | Low | High |
| Ravn‐Haren et al. (2013) | High | High | Low | Low | Low | High |
| Panetta et al. (2013) | Low | Low | High | Low | Low | Some concerns |
| Gasper et al. (2014) | High | High | Low | Low | Low | High |
| Saarenhovi et al. (2017) | High | Low | High | Low | Low | High |
| Cicero et al. (2017) | Low | Low | High | Low | Low | Some concerns |
| Tenore et al. (2018) | Low | Low | Low | Low | Low | Low |
| Gheflati et al. (2019) | Low | High | High | Low | Low | High |
| Koutsos et al. (2020) | Low | High | Low | Low | Low | Some concerns |
| Liddle et al. (2021) | Low | High | Low | Low | Low | Some concerns |
| Barnett et al. (2021) | High | High | High | Low | Low | High |
| Pushpass et al. (2023) | Low | Low | High | Low | Low | Some concerns |
| Pedret et al. (2024) | Low | High | High | Low | Low | High |
| Alzoufairi et al. (2025) | High | High | High | Low | Low | High |
Note: D1, bias arising from the randomization process; D2, bias due to deviations from the intended interventions; D3, bias due to missing outcome data; D4, bias in measurement of the outcome; D5, bias in selection of the reported result. Overall risk of bias was classified as low risk when no domains were judged to have a high risk of bias, some concerns when one domain was judged to have a high risk of bias, and high risk when more than one domain was judged to have a high risk of bias.
3.4. Meta‐Analysis
3.4.1. Effect on C‐Reactive Protein
A pooled analysis of 16 intervention arms from 13 RCTs demonstrated that apple product consumption did not significantly affect circulating CRP concentrations compared with the control groups (WMD = −0.03 mg/L, 95% CI: −0.08 to 0.03, p = 0.388). Moderate between‐study heterogeneity was observed (I 2 = 59.1%) (Figure 2A).
FIGURE 2.

Forest plot detailing weighted mean difference (WMD) and 95% confidence intervals (CIs) for the effect of apple products on: (A) CRP (mg/dL), (B) IL‐6 (pg/mL), (C) TNF‐α (pg/mL), (D) Adiponectin (mg/dL), (E) ALT (IU/L), (F) AST (IU/L), and (G) ALP (IU/L).
Exploratory subgroup analyses revealed no significant effect according to baseline CRP concentration, geographical region, sex, intervention duration, or baseline BMI. However, significant effects were observed in studies evaluating apple extract (WMD = −1.89 mg/L, p = 0.008) and apple vinegar (WMD = 0.65 mg/L, p = 0.037), although these findings should be interpreted cautiously because they were derived from a single study (Table 4).
TABLE 4.
Overall and subgroup meta‐analysis findings for the effect of apple products on serum biomarkers of inflammation, adiponectin and liver enzymes in adults.
| Number of ES | WMD (95% CI) | p | Heterogeneity | ||
|---|---|---|---|---|---|
| p heterogeneity | I 2 | ||||
| Subgroup analyses of Apple products on CRP (mg/L) | |||||
| Overall effect | 16 | −0.03 (−0.08, 0.03) | 0.388 | 0.001 | 59.1% |
| Baseline CRP (mg/L) | |||||
| > 3 | 2 | 0.03 (−0.09, 0.16) | 0.586 | 0.256 | 22.5% |
| < 3 | 5 | −0.04 (−0.10, 0.02) | 0.182 | 0.001 | 78.7% |
| Geographical region | |||||
| America | 3 | 0.14 (−0.38, 0.66) | 0.598 | 0.075 | 61.4% |
| Europe | 13 | −0.02 (−0.10, 0.04) | 0.437 | 0.004 | 59.8% |
| Other | 1 | −0.07 (−0.96, 0.82) | 0.878 | — | — |
| Sex | |||||
| Both sexes | 13 | −0.01 (−0.07, 0.04) | 0.657 | 0.075 | 38.8% |
| Female | 1 | −0.30 (−1.58, 0.98) | 0.646 | — | — |
| Male | 2 | −0.15 (−0.51, 0.21) | 0.415 | < 0.001 | 92.7% |
| Trial duration (week) | |||||
| < 8 | 8 | −0.04 (−0.11, 0.03) | 0.264 | < 0.001 | 74.8% |
| ≥ 8 | 8 | 0.07 (−0.13, 0.27) | 0.498 | 0.265 | 20.8% |
| Intervention type | |||||
| Apple Extract | 1 | −1.89 (−3.28, −0.49) | 0.008* | — | — |
| Apple Fruit | 9 | −0.01 (−0.05, 0.02) | 0.378 | 0.179 | 30.0% |
| Apple Juice | 3 | −0.24 (−0.52, 0.04) | 0.093 | 0.299 | 17.1% |
| Apple Puree | 2 | 0.03 (−0.09, 0.16) | 0.586 | 0.256 | 22.5% |
| Apple Vinegar | 1 | 0.65 (0.04, 1.26) | 0.037* | — | — |
| Baseline BMI (kg/m2) | |||||
| Normal (18.5–24.9) | 7 | −0.20 (−0.57, 0.15) | 0.262 | 0.315 | 15.1% |
| Overweight (25–29.9) | 7 | 0.01 (−0.04, 0.06) | 0.653 | 0.103 | 43.2% |
| Obese (> 30) | 2 | −0.18 (−0.47, 0.11) | 0.219 | 0.003 | 88.9% |
| Subgroup analyses of Apple products on IL‐6 (pg/mL) | |||||
| Overall effect | 5 | −0.14 (−0.31, 0.02) | 0.082 | 0.027 | 63.6% |
| Geographical region | |||||
| America | 1 | −0.33 (−0.48, −0.17) | < 0.001* | — | — |
| Europe | 4 | −0.07 (−0.22, 0.06) | 0.298 | 0.207 | 34.2% |
| Sex | |||||
| Both sexes | 4 | −0.13 (−0.30, 0.02) | 0.106 | 0.014 | 71.5% |
| Male | 1 | −0.80 (−2.64, 1.04) | 0.394 | — | — |
| Trial duration (week) | |||||
| < 8 | 3 | −0.24 (−0.38, −0.11) | < 0.001* | 0.273 | 22.9% |
| ≥ 8 | 2 | −0.00 (−0.18, 0.18) | 0.998 | 0.223 | 32.8% |
| Intervention type | |||||
| Apple Fruit | 4 | −0.13 (−0.30, 0.02) | 0.106 | 0.014 | 71.5% |
| Apple Juice | 1 | −0.80 (−2.64, 1.04) | 0.394 | — | — |
| Baseline BMI (kg/m2) | |||||
| Overweight (25–29.9) | 3 | −0.07 (−0.22, 0.08) | 0.369 | 0.136 | 49.8% |
| Obese (> 30) | 2 | −0.33 (−0.48, −0.18) | < 0.001* | 0.618 | 0.0% |
| Subgroup analyses of Apple products on TNF‐α (pg/mL) | |||||
| Overall effect | 6 | −0.09 (−0.31, 0.13) | 0.432 | 0.023 | 61.5% |
| Geographical region | |||||
| America | 1 | −0.56 (−1.06, −0.05) | 0.030* | — | — |
| Europe | 5 | 0.00 (−0.17, 0.18) | 0.963 | 0.104 | 47.9% |
| Sex | |||||
| Both sexes | 5 | −0.10 (−0.31, 0.10) | 0.332 | 0.030 | 62.5% |
| Male | 1 | 1.10 (−028, 2.48) | 0.120 | — | — |
| Trial duration (week) | |||||
| < 8 | 3 | −0.18 (−0.93, 0.56) | 0.632 | 0.088 | 58.8% |
| ≥ 8 | 3 | 0.01 (−0.14, 0.17) | 0.847 | 0.117 | 53.3% |
| Intervention | |||||
| Apple Fruit | 5 | −0.10 (−0.31, 0.10) | 0.332 | 0.030 | 62.5% |
| Apple Juice | 1 | 1.10 (−028, 2.48) | 0.120 | — | — |
| Baseline BMI (kg/m2) | |||||
| Overweight (25–29.9) | 4 | −0.00 (−0.16, 0.16) | 0.998 | 0.142 | 44.9% |
| Obese (> 30) | 2 | 0.13 (−1.46, 1.74) | 0.865 | 0.028 | 79.4% |
| Subgroup analyses of Apple products on Adiponectin (mg/dL) | |||||
| Overall effect | 3 | 0.22 (−0.35, 0.80) | 0.450 | 0.408 | 0.0% |
| Geographical region | |||||
| Asia | 1 | 0.66 (−0.30, 1.62) | 0.181 | — | — |
| Europe | 2 | −0.01 (−0.73, 0.69) | 0.960 | 0.450 | 0.0% |
| Sex | |||||
| Both sexes | 2 | 0.56 (−0.20, 1.34) | 0.152 | 0.753 | 0.0% |
| Male | 1 | −0.20 (−1.05, 0.65) | 0.647 | — | — |
| Trial duration (week) | |||||
| < 8 | 1 | −0.20 (−1.05, 0.65) | 0.647 | — | — |
| ≥ 8 | 2 | 0.56 (−0.20, 1.34) | 0.152 | 0.753 | 0.0% |
| Baseline BMI (kg/m2) | |||||
| Overweight (25–29.9) | 2 | 0.56 (−0.20, 1.34) | 0.152 | 0.753 | 0.0% |
| Obese (> 30) | 1 | −0.20 (−1.05, 0.65) | 0.647 | — | — |
| Subgroup analyses of Apple products on ALT (IU/L) | |||||
| Overall effect | 9 | −0.92 (−2.43, 0.59) | 0.234 | 0.103 | 39.8% |
| Geographical region | |||||
| Asia | 2 | −4.12 (−8.34, 0.08) | 0.055 | 0.184 | 43.4% |
| Europe | 7 | −0.00 (−1.17, 1.15) | 0.988 | 0.441 | 0.0% |
| Trial duration (week) | |||||
| < 8 | 3 | 0.24 (−2.37, 2.87) | 0.852 | 0.201 | 37.6% |
| ≥ 8 | 6 | −1.51 (−3.03, 0.00) | 0.050 | 0.357 | 9.2% |
| Intervention type | |||||
| Apple Extract | 5 | −1.39 (−4.71, 1.93) | 0.411 | 0.019 | 66.0% |
| Apple Fruit | 2 | −0.59 (−2.93, 1.74) | 0.620 | 0.551 | 0.0% |
| Apple Juice | 2 | −1.32 (−3.34, 0.69) | 0.199 | 0.843 | 0.0% |
| Baseline BMI (kg/m2) | |||||
| Normal (18.5–24.9) | 7 | −0.22 (−1.43, 0.98) | 0.716 | 0.364 | 8.5% |
| Overweight (25–29.9) | 2 | −3.91 (−10.13, 2.29) | 0.217 | 0.167 | 47.8% |
| Subgroup analyses of Apple products on AST (IU/L) | |||||
| Overall effect | 5 | −1.56 (−2.69, −0.44) | 0.006* | 0.789 | 0.0% |
| Geographical region | |||||
| Asia | 2 | −1.50 (−3.45, 0.44) | 0.131 | 0.216 | 34.6% |
| Europe | 3 | −1.57 (−3.18, 0.04) | 0.057 | 0.913 | 0.0% |
| Trial duration (week) | |||||
| < 8 | 3 | −1.25 (−2.62, 0.11) | 0.074 | 0.699 | 0.0% |
| ≥ 8 | 2 | −2.20 (−4.17, −0.24) | 0.028* | 0.537 | 0.0% |
| Baseline BMI (kg/m2) | |||||
| Normal (18.5–24.9) | 3 | −1.25 (−2.62, 0.11) | 0.074 | 0.699 | 0.0% |
| Overweight (25–29.9) | 2 | −2.20 (−4.17, −0.24) | 0.028* | 0.537 | 0.0% |
| Subgroup analyses of Apple products on ALP (IU/L) | |||||
| Overall effect | 3 | −2.81 (−6.05, 0.43) | 0.089 | 0.626 | 0.0% |
| Geographical region | |||||
| Asia | 2 | 2.38 (−8.62, 13.38) | 0.671 | 0.977 | 0.0% |
| Europe | 1 | −3.30 (−6.69, 0.09) | 0.056 | — | — |
| Trial duration (week) | |||||
| < 8 | 2 | −3.18 (−6.54, 0.17) | 0.063 | 0.635 | 0.0% |
| ≥ 8 | 1 | 2.30 (−10.00, 14.60) | 0.714 | — | — |
| Baseline BMI (kg/m2) | |||||
| Normal (18.5–24.9) | 2 | −3.18 (−6.54, 0.17) | 0.063 | 0.635 | 0.0% |
| Overweight (25–29.9) | 1 | 2.30 (−10.00, 14.60) | 0.714 | — | — |
Note: Bold values with an asterisk (*) indicate statistically significant (p < 0.05).
Abbreviations: ALP, alkaline phosphatase; ALT, alanine aminotransferase; AST, aspartate aminotransferase; CI, confidence interval; CRP, C‐reactive protein; ES, effect size; IL‐6, interleukin‐6; TNF‐α, tumor necrosis factor‐alpha; WMD, weighted mean difference.
3.4.2. Effect on Interleukin‐6
Five RCTs were included in the pooled analysis of IL‐6. Apple product supplementation did not significantly reduce circulating IL‐6 levels compared with controls (WMD = −0.14 pg/mL, 95% CI: −0.31 to 0.02, p = 0.082). Moderate heterogeneity was observed among studies (I 2 = 63.6%) (Figure 2B).
Exploratory subgroup analyses demonstrated significant reductions in studies conducted in American populations (WMD = −0.33 pg/mL, 95% CI: −0.48 to −0.17, p < 0.001), in interventions lasting < 8 weeks (WMD = −0.24 pg/mL, 95% CI: −0.38 to −0.11, p < 0.001), and among participants with obesity (BMI > 30 kg/m2) (WMD = −0.33 pg/mL, 95% CI: −0.48 to −0.18, p < 0.001). No significant effects were detected in the remaining subgroup analyses (Table 4).
3.4.3. Effect on Tumor Necrosis Factor‐Alpha
The pooled analysis of six RCTs showed that apple product consumption had no significant effect on circulating TNF‐α concentrations (WMD = −0.09 pg/mL, 95% CI: −0.31 to 0.13, p = 0.432), with moderate heterogeneity (I 2 = 61.5%) (Figure 2C). In exploratory subgroup analyses, a significant reduction was observed only in studies conducted in American populations (WMD = −0.56 pg/mL, 95% CI: −1.06 to −0.05, p = 0.030), whereas no significant effects were found according to sex, intervention duration, intervention type, or baseline BMI (Table 4).
3.4.4. Effect on Adiponectin
Three RCTs were available for adiponectin. Overall, apple products did not significantly affect circulating adiponectin concentrations (WMD = 0.22 mg/dL, 95% CI: −0.35 to 0.80, p = 0.450). No evidence of between‐study heterogeneity was observed (I 2 = 0.0%) (Figure 2D). Likewise, subgroup analyses according to geographical region, sex, intervention duration, and baseline BMI showed no statistically significant differences (Table 4).
3.4.5. Effect on Alanine Aminotransferase
Nine intervention arms from five RCTs were included in the meta‐analysis of ALT. Apple product supplementation did not significantly reduce serum ALT concentrations compared with controls (WMD = −0.92 IU/L, 95% CI: −2.43 to 0.59, p = 0.234). Low‐to‐moderate heterogeneity was observed (I 2 = 39.8%) (Figure 2E). Exploratory subgroup analyses indicated a borderline significant reduction among studies with an intervention duration of ≥ 8 weeks (WMD = −1.51 IU/L, 95% CI: −3.03 to 0.00, p = 0.050), whereas no significant differences were observed by geographical region, intervention type, or baseline BMI (Table 4).
3.4.6. Effect on Aspartate Aminotransferase
Five intervention arms from four RCTs were included in the pooled analysis of AST. Overall, apple product consumption significantly reduced serum AST concentrations (WMD = −1.56 IU/L, 95% CI: −2.69 to −0.44, p = 0.006), with no evidence of between‐study heterogeneity (I 2 = 0.0%) (Figure 2F). Exploratory subgroup analyses showed significant reductions in studies with an intervention duration of ≥ 8 weeks (WMD = −2.20 IU/L, 95% CI: −4.17 to −0.24, p = 0.028) and among overweight participants (BMI 25.0–29.9 kg/m2) (WMD = −2.20 IU/L, 95% CI: −4.17 to −0.24, p = 0.028). Although reductions were also observed in Asian and European populations and in normal‐weight participants, these subgroup effects did not reach statistical significance (Table 4).
3.4.7. Effect on Alkaline Phosphatase
Three RCTs were included in the meta‐analysis of ALP. Apple product supplementation did not significantly affect serum ALP concentrations (WMD = −2.81 IU/L, 95% CI: −6.05 to 0.43, p = 0.089), and no between‐study heterogeneity was observed (I 2 = 0.0%) (Figure 2G). Similarly, exploratory subgroup analyses by geographical region, intervention duration, and baseline BMI did not show statistically significant effects (Table 4).
3.5. Publication Bias
Publication bias was evaluated using both Begg's rank correlation test and Egger's regression asymmetry test for all investigated outcomes (Figure 3). The results indicated no evidence of publication bias for any outcome, as neither Begg's nor Egger's test reached statistical significance, as provided in Table 5. Collectively, these findings suggest a low likelihood of small‐study effects or publication bias across the included meta‐analyses.
FIGURE 3.

Funnel plots for the effect of apple products on: (A) CRP (mg/dL), (B) IL‐6 (pg/mL), (C) TNF‐α (pg/mL), (D) Adiponectin (mg/dL), (E) ALT (IU/L), (F) AST (IU/L), and (G) ALP (IU/L).
TABLE 5.
Publication bias (using Begg's and Egger's tests) and sensitivity analysis.
| Outcomes | Publication bias | Sensitivity analysis | |
|---|---|---|---|
| Begg's test | Egger's test | ||
| CRP | 0.392 | 0.730 | None |
| IL‐6 | 1.000 | 0.925 | Alzoufairi et al. 2025 (WMD: −0.21 pg/mL, 95% CI: −0.33, −0.09) |
| TNF‐α | 1.000 | 0.324 | None |
| Adiponectin | 1.000 | 0.663 | None |
| ALT | 0.917 | 0.111 | None |
| AST | 0.806 | 0.529 | Akazome et al. 2010 (A) (WMD: −1.20 IU/L, 95% CI: −2.55, 0.13) |
| ALP | 1.000 | 0.251 | None |
Abbreviations: ALP, alkaline phosphatase; ALT, alanine aminotransferase; AST, aspartate aminotransferase; CRP, C‐reactive protein; IL‐6, interleukin‐6; TNF‐α, tumor necrosis factor‐alpha.
3.6. Sensitivity Analysis
Leave‐one‐out sensitivity analyses were conducted to evaluate the robustness of the pooled estimates. The overall findings for CRP, TNF‐α, adiponectin, ALT, and ALP remained stable, indicating that no individual study had a substantial influence on the corresponding pooled effect estimates. However, the pooled estimate for IL‐6 was influenced by the exclusion of the study by Alzoufairi et al. (2025), resulting in a significant reduction in IL‐6 concentrations (WMD = −0.21 pg/mL, 95% CI: −0.33 to −0.09). Similarly, for AST, omitting Akazome et al. (2010) changed the pooled estimate to a non‐significant effect (WMD = −1.20 IU/L, 95% CI: −2.55 to 0.13) (Table 5). These findings indicate that, while most outcomes were robust to the exclusion of individual studies, the pooled effects for IL‐6 and AST were moderately influenced by single studies and should therefore be interpreted with appropriate caution.
3.7. Linear Meta‐Regression Analysis
Linear meta‐regression showed that intervention duration was significantly associated with the pooled effect of IL‐6 (coefficient = 4.79, p = 0.045), suggesting that intervention duration may contribute to between‐study variation in IL‐6 effects. No significant associations between intervention duration and the pooled effects were observed for CRP, TNF‐α, adiponectin, ALT, AST, or ALP (all p > 0.05) (Table 6 and Figure 4).
TABLE 6.
Linear meta‐regression and non‐linear duration‐response analysis (week).
| Outcomes | Linear meta‐regression | Non‐linear duration‐response | ||
|---|---|---|---|---|
| Coefficient | p | Coefficient | p | |
| CRP | −0.22 | 0.971 | 0.78 | 0.313 |
| IL‐6 | 4.79 | 0.045* | −3.45 | 0.940 |
| TNF‐α | −1.33 | 0.283 | 248.60 | 0.112 |
| Adiponectin | 8.86 | 0.136 | 20.23 | NA |
| ALT | −0.50 | 0.721 | 6.08 | 0.032* |
| AST | −1.49 | 0.601 | −8.04 | 0.202 |
| ALP | 1.56 | 0.372 | NA | NA |
Note: Non‐linear dose–response analyses were not applicable for adiponectin and ALP because of the limited number of available studies. Bold values with an asterisk (*) indicate statistically significant (p < 0.05).
Abbreviations: ALP, alkaline phosphatase; ALT, alanine aminotransferase; AST, aspartate aminotransferase; CRP, C‐reactive protein; IL‐6, interleukin‐6; NA, not available; TNF‐α, tumor necrosis factor‐alpha.
FIGURE 4.

Linear dose–response relations between Apple products and absolute mean differences. Dose–response relations between duration of intervention (week) and absolute mean differences in (A) CRP (mg/dL); (B) IL‐6 (pg/mL); (C) TNF‐α (pg/mL); (D) Adiponectin (mg/dL); (E) ALT (IU/L); (F) AST (IU/L); and (G) ALP (IU/L).
3.8. Non‐Linear Duration–Response Analysis
The non‐linear duration‐response analysis identified a significant association between intervention duration and ALT (coefficient = 6.08, p = 0.032). In contrast, no significant non‐linear duration‐response associations were observed for CRP, IL‐6, TNF‐α, or AST (all p > 0.05). Non‐linear associations could not be estimated for adiponectin or ALP because the available data were insufficient (Figure 5 and Table 6).
FIGURE 5.

Non‐linear dose–response relations between Apple products and absolute mean differences. Dose–response relations between duration of intervention (week) and absolute mean differences in (A) CRP (mg/dL); (B) IL‐6 (pg/mL); (C) TNF‐α (pg/mL); (D) Adiponectin (mg/dL); (E) ALT (IU/L); (F) AST (IU/L); and (G) ALP (IU/L).
3.9. GRADE Assessment
The GRADE assessment indicated that the certainty of evidence for the effects of apple products on inflammatory biomarkers, adiponectin, and liver enzymes ranged from Very Low to Moderate (Table 7). The certainty of evidence for CRP, IL‐6, and TNF‐α was rated Very Low, primarily due to serious concerns about risk of bias, substantial between‐study heterogeneity, and wide confidence intervals (imprecision). The certainty of evidence for adiponectin, ALT, and ALP was rated Low, with downgrading mainly attributable to risk of bias and indirectness related to differences in participant characteristics. In contrast, the certainty of evidence for AST was rated Moderate, with the evidence downgraded only for risk of bias. Overall, the GRADE assessment indicates limited confidence in the estimated effects of apple products for most inflammatory and hepatic biomarkers.
TABLE 7.
GRADE certainty of evidence profile of the effect of apple products on serum biomarkers of inflammation, adiponectin, and liver enzymes in adults.
| Outcomes | Risk of bias | Inconsistency | Indirectness | Imprecision | Publication bias | Quality of evidence |
|---|---|---|---|---|---|---|
| CRP | Serious limitation a | Serious limitation b | No serious limitation | Serious limitation c | No serious limitation |
⊕◯◯◯ Very Low |
| IL‐6 | Serious limitation a | Serious limitation b | No serious limitation | Serious limitation c | No serious limitation |
⊕◯◯◯ Very Low |
| TNF‐α | Serious limitation a | Serious limitation b | No serious limitation | Serious limitation c | No serious limitation |
⊕◯◯◯ Very Low |
| Adiponectin | Serious limitation a | No serious limitation | No serious limitation | Serious limitation c | No serious limitation |
⊕⊕◯◯ Low |
| ALT | Serious limitation a | No serious limitation | No serious limitation | Serious limitation c | No serious limitation |
⊕⊕◯◯ Low |
| AST | Serious limitation a | No serious limitation | No serious limitation | No serious limitation | No serious limitation |
⊕⊕⊕◯ Moderate |
| ALP | Serious limitation a | No serious limitation | No serious limitation | Serious limitation c | No serious limitation |
⊕⊕◯◯ Low |
Abbreviations: ALP, alkaline phosphatase; ALT, alanine aminotransferase; AST, aspartate aminotransferase; CRP, C‐reactive protein; IL‐6, interleukin‐6; TNF‐α, tumor necrosis factor‐alpha.
Most of the included studies have a high risk of bias.
There is a moderate level of heterogeneity (I 2 > 40%).
Participants vary in terms of clinical conditions, sex, or age group.
4. Discussion
4.1. Aim and Overall Findings
This systematic review and meta‐analysis examined the effects of a wide range of apple‐derived products and found no overall significant changes in markers such as CRP, IL‐6, TNF‐α, adiponectin, ALT, or ALP. However, several exploratory subgroup analyses suggested that the effects may vary depending on participant characteristics and intervention type. The clearest and most consistent result was a reduction in serum AST levels, an effect that was significant among trials with an intervention duration of ≥ 8 weeks, whereas the reduction observed in shorter trials did not reach statistical significance. It appeared to be more evident among Asian participants, middle‐aged or overweight individuals, and studies using apple extract preparations. Although reductions in IL‐6 and TNF‐α were observed in selected analyses, the corresponding estimates were based on a limited number of contributing trials [insert relevant study citations] and should therefore be interpreted cautiously. A modest duration‐related effect was observed for ALT, but the association did not reach strong statistical significance. In contrast, adiponectin and ALP did not appear to change meaningfully in any subgroup analysis.
The observed effects should be interpreted as responses to the respective apple‐derived interventions rather than as evidence of isolated polyphenol activity. Whole apples contain a complex food matrix comprising polyphenols, dietary fiber, organic acids, micronutrients, and other bioactive constituents that may act independently or synergistically. In addition, processing into extracts, juices, purées, or vinegar can substantially modify the composition, bioavailability, and relative contribution of these components. Therefore, differences observed across intervention types should not be interpreted simply as differences in polyphenol exposure. Alternative mechanisms may include effects related to dietary fiber, particularly pectin, gastrointestinal fermentation and gut microbial metabolism, organic acids such as acetic acid in vinegar, or interactions among multiple components of the apple matrix. Evidence from dietary‐fiber interventions suggests that fiber may influence selected inflammatory biomarkers, although effects are inconsistent across outcomes and populations; thus, such evidence provides a plausible alternative contributor rather than establishing that fiber mediates the effects observed in the present analysis.
The heterogeneity of apple‐derived interventions is also relevant to the interpretation of the pooled findings. Whole apples, extracts, juices, purées, and apple vinegar differ substantially in their composition and processing, including the relative concentrations and bioavailability of polyphenols, dietary fiber, and organic acids. Therefore, these interventions should not be interpreted as equivalent forms of polyphenol exposure. Components other than polyphenols, as well as interactions among constituents within the apple matrix, may contribute to the observed effects. Evidence from dietary‐fiber interventions provides a plausible context for such alternative mechanisms; however, it does not establish that fiber mediates the effects observed in the present analysis.
A recent systematic review by Del Bo' et al. (2026) evaluated apples and apple‐based products across a broad range of cardiometabolic and functional outcomes, including inflammatory biomarkers. The present review extends this evidence by quantitatively synthesizing data on predefined inflammatory biomarkers, adiponectin, and liver enzymes, using evidence exclusively from randomized controlled trials, while also incorporating heterogeneous interventions such as apple cider vinegar and extracts, along with an updated literature search. The overlap between the two evidence bases is therefore expected; however, the present analysis provides a more outcome‐specific quantitative synthesis of inflammatory, adipokine, and hepatic biomarkers and explicitly examines potential sources of heterogeneity across different forms of apple‐derived interventions.
4.2. Possible Physiological Mechanism
The biological effects of apple‐derived products on inflammatory and hepatic biomarkers observed in this meta‐analysis are primarily attributable to the rich and diverse polyphenolic composition of apple products, which includes procyanidins as the predominant fraction (approximately 45% of total polyphenols), alongside quercetin glycosides, chlorogenic acid, catechin and related flavan‐3‐ol monomers, and phloridzin (Yoshida et al. 2023), with apple vinegar additionally contributing acetic acid as a biologically active component (Chiu et al. 2020). Apple polyphenols, especially procyanidins, are known to exert anti‐inflammatory effects by suppressing key pro‐inflammatory factors, including NF‐κB and MAPK. Through this suppression, IkBa phosphorylation is inhibited, nuclear translocation of NF‐κB/p65 is prevented, and the expression and release of cytokines such as TNF‐α and IL‐6 are reduced. Consequently, levels of acute‐phase reactants, including CRP, are decrease (Han et al. 2019; Wang et al. 2020).
The greater responsiveness of obese and overweight individuals to IL‐6 and AST reductions is consistent with adiposity‐driven chronic low‐grade inflammation; hypertrophic adipose tissue triggers macrophage M1 polarization and NF‐κB‐mediated cytokine release, whose portal drainage into the liver promotes hepatic lipid peroxidation and transaminase elevation (Khanna et al. 2022; Saltiel and Olefsky 2017). These observations suggest that overweight and obese individuals may represent the population most likely to derive meaningful benefit from apple‐derived bioactive compounds. However, confirmation in trials specifically targeting these populations is warranted.
The hepatoprotective signal observed for AST in this analysis, with apple polyphenol extracts as the primary driver, is underpinned by procyanidin‐mediated promotion of acyl‐CoA oxidation enzyme activity in the liver alongside concurrent suppression of fatty acid synthase, collectively shifting hepatic metabolism away from lipogenic and steatotic pathways that damage hepatocyte integrity (Nagasako‐Akazome et al. 2007; Pour et al. 2025). This mechanism is further explained by direct clinical measurement of hepatic enzyme profiles in patients with type 2 diabetes mellitus receiving apple vinegar intervention, which yielded no significant change in AST, ALT, or ALP, consistent with the absence of meaningful visceral fat reduction in that population, suggesting that the effects of the apple‐derived intervention on transaminase levels may depend on the underlying metabolic context; however, the available evidence does not establish a specific polyphenol‐mediated mechanism (Mahmoodi et al. 2013). Annurca apple procyanidins have further demonstrated the capacity to enhance apolipoprotein A1 expression and to favor LDL receptor binding activity in hepatocytes, thereby improving intracellular lipid clearance and reducing hepatocellular stress (Tenore et al. 2018). For vinegar‐based apple products, the hepatic and metabolic mechanisms are additionally mediated by acetic acid through AMPK activation, suppression of SREBP‐1c, and inhibition of fatty acid synthase (Li et al. 2013), pathways that reduce lipogenesis, attenuate glucotoxicity‐driven hepatic inflammation, as evidenced by metabolic improvements documented in clinical trials of apple cider vinegar (Jafarirad et al. 2023; Khezri et al. 2018; Kondo et al. 2009). The consistent failure of apple products to produce a significant change in adiponectin across all included studies is consistent with the currently accepted understanding that adiponectin secretion from fat cells (adipocytes) is strongly dependent on PPARγ activation and a reduction in adipose tissue mass, suggesting that the observed effects may depend on the characteristics of the apple‐derived intervention and the metabolic context of the participants, particularly in the absence of substantial caloric restriction or weight loss (Nagasako‐Akazome et al. 2007; Astapova and Leff 2012).
4.3. Strengths and Limitations
This review has several methodological and analytical strengths. The exclusive inclusion of randomized controlled trials strengthens the evidentiary basis of the quantitative synthesis, while predefined outcome‐specific meta‐analyses enabled direct estimation of the effects of apple‐derived interventions on inflammatory biomarkers, adiponectin, and liver enzymes. The inclusion of different apple‐derived intervention forms allowed potential differences according to product type to be explored, and predefined subgroup and meta‐regression analyses were used to investigate potential sources of between‐study heterogeneity. Risk of bias was assessed using an established Cochrane framework, and the certainty of evidence was evaluated using GRADE, supporting transparent interpretation of the pooled findings.
Several limitations should be considered when interpreting the findings of this meta‐analysis. Although only RCTs were included, many studies were judged to have a high risk of bias, particularly in the domains of randomization and deviations from intended interventions, which may reduce confidence in some pooled estimates. In addition, considerable clinical and methodological heterogeneity existed across studies regarding participant characteristics, intervention duration, dosage, and the type of apple‐derived product evaluated, including whole apples, juices, polyphenol extracts, pomace, purees, and apple cider vinegar, each differing in bioactive composition and bioavailability. Several outcomes, particularly adiponectin, AST, and ALP, were assessed in a limited number of trials, restricting statistical power and the reliability of subgroup and sensitivity analyses. Furthermore, most interventions were relatively short in duration, limiting conclusions about the long‐term effects of apple‐derived products on inflammatory and hepatic biomarkers. In addition, an additional limitation is the heterogeneity of the apple‐derived interventions. Whole apples, extracts, juices, purées, and apple vinegar differ in composition, processing, dose, and bioavailability of their constituent compounds. Consequently, pooled estimates should be interpreted as effects of the respective apple‐derived interventions rather than as evidence for the effect of a single bioactive component. Finally, potentially important modifiers of treatment response, such as habitual dietary intake, physical activity, medication use, gut microbiota composition, and inter‐individual differences in polyphenol metabolism, were inconsistently reported and could not be adequately evaluated. Therefore, the present findings should be interpreted with appropriate caution and confirmed by larger, high‐quality, long‐term RCTs using standardized apple‐derived interventions.
4.4. Future Research Direction
Future research should prioritize adequately powered, well‐designed RCTs using standardized apple products with quantified polyphenol and other bioactive compound content, longer intervention durations, and clearly defined doses to establish dose–response relationships and long‐term efficacy. Direct comparisons of whole apples, extracts, juices, pomace, and apple cider vinegar may help clarify the relative contributions of polyphenols, fiber, and acetic acid. Studies should also investigate sources of inter‐individual variability by considering metabolic phenotype, geographical region, and gut microbiota composition, particularly among populations with obesity, metabolic dysfunction, non‐alcoholic fatty liver disease, or elevated inflammatory biomarkers. Finally, mechanistic studies assessing pathways such as NF‐κB and Nrf2, together with gut microbiota–mediated polyphenol metabolism, may help explain heterogeneous clinical responses and strengthen the translation of mechanistic findings into clinical applications.
5. Conclusion
Apple and apple‐derived products did not significantly improve most evaluated inflammatory biomarkers, adiponectin, or liver enzymes in adults, although a significant reduction in AST was observed. Exploratory subgroup analyses suggested that effects may vary by apple product type, intervention duration, and participant characteristics. Given the low to very low certainty of evidence for most outcomes, these findings should be interpreted cautiously. Further well‐designed, adequately powered RCTs with standardized apple interventions and longer follow‐up are warranted to clarify their potential effects on inflammation, adipokine regulation, and hepatic biomarkers.
Author Contributions
Sahel Hamednia: methodology, software, data curation, investigation, validation, funding acquisition, visualization, resources. Mehdi Karimi: conceptualization, methodology, software, investigation, validation, supervision, funding acquisition, visualization, project administration, resources, writing – original draft, writing – review and editing. Camellia Akhgarjand: conceptualization, methodology, software, investigation, validation, funding acquisition, visualization, project administration, resources, writing – original draft. Elaheh Dehghani: methodology, software, data curation, investigation, validation, funding acquisition, visualization, resources. Melina Todehkharman: writing – review and editing, writing – original draft, funding acquisition, conceptualization, software, resources, validation, investigation. Mohammad Mehdi Khaleghi: methodology, software, data curation, investigation, validation, funding acquisition, visualization, resources. Omid Asbaghi: conceptualization, methodology, software, data curation, investigation, validation, formal analysis, supervision, funding acquisition, visualization, project administration, resources, writing – review and editing. Ali Hosseini: supervision, project administration, methodology, investigation, validation. Moslem Naderian: methodology, validation, supervision, funding acquisition, project administration.
Funding
This work did not receive any specific grant from funding agencies in the public, commercial, or not‐for‐profit sectors.
Ethics Statement
The authors have nothing to report.
Consent
The authors have nothing to report.
Conflicts of Interest
The authors declare no conflicts of interest.
Supporting information
Supplementary Table 1. Search keywords and search strategy.
Supplementary Table 2. Query boxes of each database based on the Title & Abstract (search date: August 1, 2026).
Acknowledgments
Declaration of AI‐assisted tools: The authors used ChatGPT (OpenAI) to assist with generating and designing the graphical abstract presented in this manuscript. The authors reviewed, verified, and approved the content and final appearance of the figure and remain fully responsible for its accuracy, interpretation, and scientific integrity. ChatGPT was not used to generate, analyze, or interpret the study data. Grammarly was used solely to assist with English‐language grammar, spelling, and stylistic editing. All scientific content, interpretations, and final manuscript text were reviewed and approved by the authors.
Contributor Information
Mehdi Karimi, Email: karimi9010@gmail.com.
Ali Hosseini, Email: alihosseini13@yahoo.com.
Moslem Naderian, Email: m_naderian_66@yahoo.com.
Data Availability Statement
All data used in this meta‐analysis were extracted from published studies. The datasets supporting the findings of this study are available from the original sources, which are cited in the manuscript. Additional information can be provided by the corresponding author upon reasonable request.
References
- Akazome, Y. , Kametani N., Kanda T., Shimasaki H., and Kobayashi S.. 2010. “Evaluation of Safety of Excessive Intake and Efficacy of Long‐Term Intake of Beverages Containing Apple Polyphenols.” Journal of Oleo Science 59, no. 6: 321–338. 10.5650/jos.59.321. [DOI] [PubMed] [Google Scholar]
- Alzoufairi, S. , Pushpass R. G., Liu L., Lovegrove J. A., and Jackson K. G.. 2025. “Impact of Chronic Consumption of Probiotics, Oats, and Apples on Expression of Genes Related to Bile Acids, Lipid, Gut Peptides, and Inflammation in Peripheral Monocular Cells—Findings From the CABALA Study.” European Journal of Nutrition 64, no. 4: 176. 10.1007/s00394-025-03694-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Amir‐Behghadami, M. , and Janati A.. 2020. “Population, Intervention, Comparison, Outcomes and Study (PICOS) Design as a Framework to Formulate Eligibility Criteria in Systematic Reviews.” Emergency Medicine Journal 37: 387. 10.1136/emermed-2020-209567. [DOI] [PubMed] [Google Scholar]
- Astapova, O. , and Leff T.. 2012. “Adiponectin and PPARγ: Cooperative and Interdependent Actions of Two Key Regulators of Metabolism.” Vitamins and Hormones 90: 143–162. 10.1016/B978-0-12-398313-8.00006-3. [DOI] [PubMed] [Google Scholar]
- Auclair, S. , Chironi G., Milenkovic D., et al. 2010. “The Regular Consumption of a Polyphenol‐Rich Apple Does Not Influence Endothelial Function: A Randomised Double‐Blind Trial in Hypercholesterolemic Adults.” European Journal of Clinical Nutrition 64, no. 10: 1158–1165. 10.1038/ejcn.2010.135. [DOI] [PubMed] [Google Scholar]
- Barnett, M. P. G. , Young W., Armstrong K., et al. 2021. “A Polyphenol Enriched Variety of Apple Alters Circulating Immune Cell Gene Expression and Faecal Microbiota Composition in Healthy Adults: A Randomized Controlled Trial.” Nutrients 13, no. 4: 1092. 10.3390/nu13041092. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Bars‐Cortina, D. , Martínez‐Bardají A., Macià A., Motilva M. J., and Piñol‐Felis C.. 2020. “Consumption Evaluation of One Apple Flesh a Day in the Initial Phases Prior to Adenoma/Adenocarcinoma in an Azoxymethane Rat Colon Carcinogenesis Model.” Journal of Nutritional Biochemistry 83: 108418. 10.1016/j.jnutbio.2020.108418. [DOI] [PubMed] [Google Scholar]
- Barth, S. W. , Koch T. C., Watzl B., Dietrich H., Will F., and Bub A.. 2012. “Moderate Effects of Apple Juice Consumption on Obesity‐Related Markers in Obese Men: Impact of Diet‐Gene Interaction on Body Fat Content.” European Journal of Nutrition 51, no. 7: 841–850. 10.1007/s00394-011-0264-6. [DOI] [PubMed] [Google Scholar]
- Begg, C. B. , and Mazumdar M.. 1994. “Operating Characteristics of a Rank Correlation Test for Publication Bias.” Biometrics 50, no. 4: 1088–1101. 10.2307/2533446. [DOI] [PubMed] [Google Scholar]
- Bondonno, N. P. , Bondonno C. P., Blekkenhorst L. C., et al. 2018. “Flavonoid‐Rich Apple Improves Endothelial Function in Individuals at Risk for Cardiovascular Disease: A Randomized Controlled Clinical Trial.” Molecular Nutrition & Food Research 62, no. 3. 10.1002/mnfr.201700674. [DOI] [PubMed] [Google Scholar]
- Borenstein, M. , L., V. Hedges , J. P. T. Higgins , and Rothstein H. R.. 2009. “How a Meta‐Analysis Works.” In Introduction to Meta‐Analysis, 1–7. John Wiley & Sons Ltd. https://onlinelibrary.wiley.com/doi/book/10.1002/9780470743386. [Google Scholar]
- Chai, S. C. , Hooshmand S., Saadat R. L., Payton M. E., Brummel‐Smith K., and Arjmandi B. H.. 2012. “Daily Apple Versus Dried Plum: Impact on Cardiovascular Disease Risk Factors in Postmenopausal Women.” Journal of the Academy of Nutrition and Dietetics 112, no. 8: 1158–1168. 10.1016/j.jand.2012.05.005. [DOI] [PubMed] [Google Scholar]
- Chiu, H.‐F. , Chiang M., Liao H.‐J., et al. 2020. “The Ergogenic Activity of Cider Vinegar: A Randomized Cross‐Over, Double‐Blind, Clinical Trial.” Sports Medicine and Health Science 2, no. 1: 38–43. 10.1016/j.smhs.2020.02.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Cicero, A. F. G. , Caliceti C., Fogacci F., et al. 2017. “Effect of Apple Polyphenols on Vascular Oxidative Stress and Endothelium Function: A Translational Study.” Molecular Nutrition & Food Research 61, no. 11. 10.1002/mnfr.201700373. [DOI] [PubMed] [Google Scholar]
- Cifuentes, M. , Verdejo H. E., Castro P. F., et al. 2024. “Low‐Grade Chronic Inflammation: A Shared Mechanism for Chronic Diseases.” Physiology 40, no. 1: 4–25. 10.1152/physiol.00021.2024. [DOI] [PubMed] [Google Scholar]
- Clemente‐Suárez, V. J. , Redondo‐Flórez L., Beltrán‐Velasco A. I., et al. 2023. “The Role of Adipokines in Health and Disease.” Biomedicine [Internet] 11, no. 5: 1290. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Del Bo', C. , Martini D., Giancristofaro A., et al. 2026. “Apples and Apple‐Based Products in the Modulation of Cardiometabolic and Functional Markers: A Systematic Review of Human Intervention Studies.” Food & Function 17, no. 13: 5858–5879. [DOI] [PubMed] [Google Scholar]
- Del Bo, C. , Martini D., Porrini M., Klimis‐Zacas D., and Riso P.. 2015. “Berries and Oxidative Stress Markers: An Overview of Human Intervention Studies.” Food & Function 6, no. 9: 2890–2917. [DOI] [PubMed] [Google Scholar]
- DerSimonian, R. , and Laird N.. 1986. “Meta‐Analysis in Clinical Trials.” Controlled Clinical Trials 7, no. 3: 177–188. 10.1016/0197-2456(86)90046-2. [DOI] [PubMed] [Google Scholar]
- Egger, M. , Davey Smith G., Schneider M., and Minder C.. 1997. “Bias in Meta‐Analysis Detected by a Simple, Graphical Test.” BMJ 315, no. 7109: 629–634. 10.1136/bmj.315.7109.629. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Garmo, G. R. , Zhang Z., Ofgea G. C., Ruojun D., Kuffi K. D., and Zhang J.. 2026. “Organic and Conventional Farming Systems: Effects on Fruit Quality and Nutritional Attributes of Apple ( Malus domestica Borkh.)—A Review.” Agricultural Systems 236: 104769. 10.1016/j.agsy.2026.104769. [DOI] [Google Scholar]
- Gasper, A. , Hollands W., Casgrain A., et al. 2014. “Consumption of Both Low and High (−)‐Epicatechin Apple Puree Attenuates Platelet Reactivity and Increases Plasma Concentrations of Nitric Oxide Metabolites: A Randomized Controlled Trial.” Archives of Biochemistry and Biophysics 559: 29–37. 10.1016/j.abb.2014.05.026. [DOI] [PubMed] [Google Scholar]
- Gheflati, A. , Bashiri R., Ghadiri‐Anari A., Reza J. Z., Kord M. T., and Nadjarzadeh A.. 2019. “The Effect of Apple Vinegar Consumption on Glycemic Indices, Blood Pressure, Oxidative Stress, and Homocysteine in Patients With Type 2 Diabetes and Dyslipidemia: A Randomized Controlled Clinical Trial.” Clinical Nutrition ESPEN 33: 132–138. 10.1016/j.clnesp.2019.06.006. [DOI] [PubMed] [Google Scholar]
- Guo, C. , Wei J., Yang J., Xu J., Pang W., and Jiang Y.. 2008. “Pomegranate Juice Is Potentially Better Than Apple Juice in Improving Antioxidant Function in Elderly Subjects.” Nutrition Research 28, no. 2: 72–77. 10.1016/j.nutres.2007.12.001. [DOI] [PubMed] [Google Scholar]
- Guyatt, G. H. , Oxman A. D., Vist G. E., et al. 2008. “GRADE: An Emerging Consensus on Rating Quality of Evidence and Strength of Recommendations.” BMJ 336, no. 7650: 924–926. 10.1136/bmj.39489.470347.AD. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Han, S. , Gao H., Chen S., et al. 2019. “Procyanidin A1 Alleviates Inflammatory Response Induced by LPS Through NF‐κB, MAPK, and Nrf2/HO‐1 Pathways in RAW264. 7 Cells.” Scientific Reports 9, no. 1: 15087. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Higgins, J. , Thomas J., Chandler J., et al. 2019. Cochrane Handbook for Systematic Reviews of Interventions. Wiley. 10.1002/9781119536604. [DOI] [Google Scholar]
- Hodgson, J. M. , Prince R. L., Woodman R. J., et al. 2016. “Apple Intake Is Inversely Associated With All‐Cause and Disease‐Specific Mortality in Elderly Women.” British Journal of Nutrition 115, no. 5: 860–867. 10.1017/S0007114515005231. [DOI] [PubMed] [Google Scholar]
- Hollands, W. J. , Tapp H., Defernez M., et al. 2018. “Lack of Acute or Chronic Effects of Epicatechin‐Rich and Procyanidin‐Rich Apple Extracts on Blood Pressure and Cardiometabolic Biomarkers in Adults With Moderately Elevated Blood Pressure: A Randomized, Placebo‐Controlled Crossover Trial.” American Journal of Clinical Nutrition 108, no. 5: 1006–1014. 10.1093/ajcn/nqy139. [DOI] [PubMed] [Google Scholar]
- Hozo, S. P. , Djulbegovic B., and Hozo I.. 2005. “Estimating the Mean and Variance From the Median, Range, and the Size of a Sample.” BMC Medical Research Methodology 5: 13. 10.1186/1471-2288-5-13. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Ioannidis, J. P. 2008. “Interpretation of Tests of Heterogeneity and Bias in Meta‐Analysis.” Journal of Evaluation in Clinical Practice 14, no. 5: 951–957. 10.1111/j.1365-2753.2008.00986.x. [DOI] [PubMed] [Google Scholar]
- Jafarirad, S. , Elahi M.‐R., Mansoori A., Khanzadeh A., and Haghighizadeh M.‐H.. 2023. “The Improvement Effect of Apple Cider Vinegar as a Functional Food on Anthropometric Indices, Blood Glucose and Lipid Profile in Diabetic Patients: A Randomized Controlled Clinical Trial.” Frontiers in Clinical Diabetes and Healthcare 4: 1288786. 10.3389/fcdhc.2023.1288786. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Joseph, S. V. , Edirisinghe I., and Burton‐Freeman B. M.. 2016. “Fruit Polyphenols: A Review of Anti‐Inflammatory Effects in Humans.” Critical Reviews in Food Science and Nutrition 56, no. 3: 419–444. 10.1080/10408398.2013.767221. [DOI] [PubMed] [Google Scholar]
- Karimi, M. , Abidi M., Gholamian R., et al. 2026. “Vitamin E Administration Reduces Liver Enzyme Levels in Patients With Metabolic Dysfunction‐Associated Steatotic Liver Disease (MASLD): Evidence From a Systematic Review and Meta‐Analysis.” Nutrition & Metabolism 23, no. 1: 105. 10.1186/s12986-026-01158-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Karimi, M. , Javadi M., Hamzavi S. F., et al. 2025. “Effects of Chokeberry Products on Cardiometabolic Indices in Adults: A GRADE‐Assessed Systematic Review and Meta‐Analysis of RCTs.” Nutrire 50, no. 2: 53. 10.1186/s41110-025-00359-x. [DOI] [Google Scholar]
- Karimi, M. , Javadi M., Sharifi M., Valizadeh F., Karimi M. A., and Asbaghi O.. 2025. “Effects of Curcuminoids Plus Piperine Co‐Supplementation on Liver Enzymes and Inflammation in Adults: A GRADE‐Assessed Systematic Review and Meta‐Analysis.” Food Science & Nutrition 13, no. 7: e70588. 10.1002/fsn3.70588. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Khanna, D. , Khanna S., Khanna P., Kahar P., and Patel B. M.. 2022. “Obesity: A Chronic Low‐Grade Inflammation and Its Markers.” Cureus 14, no. 2. 10.7759/cureus.22711. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Khezri, S. S. , Saidpour A., Hosseinzadeh N., and Amiri Z.. 2018. “Beneficial Effects of Apple Cider Vinegar on Weight Management, Visceral Adiposity Index and Lipid Profile in Overweight or Obese Subjects Receiving Restricted Calorie Diet: A Randomized Clinical Trial.” Journal of Functional Foods 43: 95–102. 10.1016/j.jff.2018.02.003. [DOI] [Google Scholar]
- Kirichenko, T. V. , Markina Y. V., Bogatyreva A. I., Tolstik T. V., Varaeva Y. R., and Starodubova A. V.. 2022. “The Role of Adipokines in Inflammatory Mechanisms of Obesity.” International Journal of Molecular Sciences 23, no. 23: 14982. 10.3390/ijms232314982. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Kondo, T. , Kishi M., Fushimi T., Ugajin S., and Kaga T.. 2009. “Vinegar Intake Reduces Body Weight, Body Fat Mass, and Serum Triglyceride Levels in Obese Japanese Subjects.” Bioscience, Biotechnology, and Biochemistry 73, no. 8: 1837–1843. 10.1271/bbb.90231. [DOI] [PubMed] [Google Scholar]
- Koutsos, A. , Lima M., Conterno L., et al. 2017. “Effects of Commercial Apple Varieties on Human Gut Microbiota Composition and Metabolic Output Using an In Vitro Colonic Model.” Nutrients 9, no. 6: 533. 10.3390/nu9060533. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Koutsos, A. , Riccadonna S., Ulaszewska M. M., et al. 2020. “Two Apples a Day Lower Serum Cholesterol and Improve Cardiometabolic Biomarkers in Mildly Hypercholesterolemic Adults: A Randomized, Controlled, Crossover Trial.” American Journal of Clinical Nutrition 111, no. 2: 307–318. 10.1093/ajcn/nqz282. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Koutsos, A. , Tuohy K. M., and Lovegrove J. A.. 2015. “Apples and Cardiovascular Health—Is the Gut Microbiota a Core Consideration?” Nutrients 7, no. 6: 3959–3998. 10.3390/nu7063959. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Lee, K. W. , Kim Y. J., Kim D. O., Lee H. J., and Lee C. Y.. 2003. “Major Phenolics in Apple and Their Contribution to the Total Antioxidant Capacity.” Journal of Agricultural and Food Chemistry 51, no. 22: 6516–6520. 10.1021/jf034475w. [DOI] [PubMed] [Google Scholar]
- Li, X. , Chen H., Guan Y., et al. 2013. “Acetic Acid Activates the AMP‐Activated Protein Kinase Signaling Pathway to Regulate Lipid Metabolism in Bovine Hepatocytes.” PLoS One 8, no. 7: e67880. 10.1371/journal.pone.0067880. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Liddle, D. M. , Lin X., Cox L. C., et al. 2021. “Daily Apple Consumption Reduces Plasma and Peripheral Blood Mononuclear Cell‐Secreted Inflammatory Biomarkers in Adults With Overweight and Obesity: A 6‐Week Randomized, Controlled, Parallel‐Arm Trial.” American Journal of Clinical Nutrition 114, no. 2: 752–763. 10.1093/ajcn/nqab094. [DOI] [PubMed] [Google Scholar]
- Liu, C. , Shao M., Lu L., Zhao C., Qiu L., and Liu Z.. 2021. “Obesity, Insulin Resistance and Their Interaction on Liver Enzymes.” PLoS One 16, no. 4: e0249299. 10.1371/journal.pone.0249299. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Mahmoodi, M. , Hosseini‐zijoud S.‐M., Hassanshahi G., et al. 2013. “The Effect of White Vinegar on Some Blood Biochemical Factors in Type 2 Diabetic Patients.” Journal of Diabetes and Endocrinology 4, no. 1: 1–5. [Google Scholar]
- Nagasako‐Akazome, Y. , Kanda T., Ohtake Y., Shimasaki H., and Kobayashi T.. 2007. “Apple Polyphenols Influence Cholesterol Metabolism in Healthy Subjects With Relatively High Body Mass Index.” Journal of Oleo Science 56, no. 8: 417–428. 10.5650/jos.56.417. [DOI] [PubMed] [Google Scholar]
- Page, M. J. , McKenzie J. E., Bossuyt P. M., et al. 2021. “The PRISMA 2020 Statement: An Updated Guideline for Reporting Systematic Reviews.” BMJ 372: n71. 10.1136/bmj.n71. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Panetta, C. , Jonk Y., and Shapiro A.. 2013. “Prospective Randomized Clinical Trial Evaluating the Impact of Vinegar on Lipids in Non‐Diabetics.” World Journal of Cardiovascular Diseases 03: 191–196. 10.4236/wjcd.2013.32027. [DOI] [Google Scholar]
- Pedret, A. , Companys J., Calderón‐Pérez L., et al. 2024. “A Red‐Fleshed Apple Rich in Anthocyanins Improves Endothelial Function, Reduces Inflammation, and Modulates the Immune System in Hypercholesterolemic Subjects: The AppleCOR Study.” Food & Function 15, no. 11: 5825–5841. 10.1039/D3FO05114E. [DOI] [PubMed] [Google Scholar]
- Pour, P. H. , Hasanpour M., Kesharwani P., Sobhani Z., and Sahebkar A.. 2025. “Exploring the Hepatoprotective Effects of Apples: A Comprehensive Review of Bioactive Compounds and Molecular Mechanisms.” Fitoterapia 185: 106689. 10.1016/j.fitote.2025.106689. [DOI] [PubMed] [Google Scholar]
- Pushpass, R. G. , Alzoufairi S., Mancini A., et al. 2023. “Chronic Consumption of Probiotics, Oats, and Apples Has Differential Effects on Postprandial Bile Acid Profile and Cardiometabolic Disease Risk Markers Compared With an Isocaloric Control (Cornflakes): A Randomized Trial.” American Journal of Clinical Nutrition 117, no. 2: 252–265. 10.1016/j.ajcnut.2022.10.013. [DOI] [PubMed] [Google Scholar]
- Ravn‐Haren, G. , Dragsted L. O., Buch‐Andersen T., et al. 2013. “Intake of Whole Apples or Clear Apple Juice Has Contrasting Effects on Plasma Lipids in Healthy Volunteers.” European Journal of Nutrition 52, no. 8: 1875–1889. 10.1007/s00394-012-0489-z. [DOI] [PubMed] [Google Scholar]
- Ruan, W. , Zhou X., Liu H., Wang T., Zhang G., and Lin K.. 2024. “Causal Role of Circulating Inflammatory Cytokines in Cardiac Diseases, Structure and Function.” Heart & Lung 67: 70–79. 10.1016/j.hrtlng.2024.04.018. [DOI] [PubMed] [Google Scholar]
- Saarenhovi, M. , Salo P., Scheinin M., et al. 2017. “The Effect of an Apple Polyphenol Extract Rich in Epicatechin and Flavan‐3‐Ol Oligomers on Brachial Artery Flow‐Mediated Vasodilatory Function in Volunteers With Elevated Blood Pressure.” Nutrition Journal 16, no. 1: 73. 10.1186/s12937-017-0291-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Saltiel, A. R. , and Olefsky J. M.. 2017. “Inflammatory Mechanisms Linking Obesity and Metabolic Disease.” Journal of Clinical Investigation 127, no. 1: 1–4. 10.1172/JCI92035. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Sandoval‐Ramírez, B. A. , Catalán Ú., Calderón‐Pérez L., et al. 2020. “The Effects and Associations of Whole‐Apple Intake on Diverse Cardiovascular Risk Factors. A Narrative Review.” Critical Reviews in Food Science and Nutrition 60, no. 22: 3862–3875. 10.1080/10408398.2019.1709801. [DOI] [PubMed] [Google Scholar]
- Setayesh, A. , Karimi M., Valizadeh F., et al. 2025. “Probiotics and Synbiotics Supplementation Reduce Inflammatory Cytokines in Individuals With Prediabetes and Type 2 Diabetes Mellitus: Findings From a Systematic Review Meta‐Analysis.” Advances in Nutrition 16, no. 11: 100526. 10.1016/j.advnut.2025.100526. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Shi, L. , and Lin L.. 2019. “The Trim‐And‐Fill Method for Publication Bias: Practical Guidelines and Recommendations Based on a Large Database of Meta‐Analyses.” Medicine (Baltimore) 98, no. 23: e15987. 10.1097/MD.0000000000015987. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Sterne, J. A. C. , Savović J., Page M. J., et al. 2019. “RoB 2: A Revised Tool for Assessing Risk of Bias in Randomised Trials.” BMJ 366: l4898. 10.1136/bmj.l4898. [DOI] [PubMed] [Google Scholar]
- Tenore, G. C. , Carotenuto A., Caruso D., et al. 2018. “A Nutraceutical Formulation Based on Annurca Apple Polyphenolic Extract Is Effective on Intestinal Cholesterol Absorption: A Randomised, Placebo‐Controlled, Crossover Study.” PharmaNutrition 6: 85–94. [Google Scholar]
- Wang, Q.‐Q. , Gao H., Yuan R., et al. 2020. “Procyanidin A2, a Polyphenolic Compound, Exerts Anti‐Inflammatory and Anti‐Oxidative Activity in Lipopolysaccharide‐Stimulated RAW264. 7 Cells.” PLoS One 15, no. 8: e0237017. 10.1371/journal.pone.0237017. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Wang, X. , Ouyang Y., Liu J., et al. 2014. “Fruit and Vegetable Consumption and Mortality From All Causes, Cardiovascular Disease, and Cancer: Systematic Review and Dose‐Response Meta‐Analysis of Prospective Cohort Studies.” BMJ 349: g4490. 10.1136/bmj.g4490. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Xie, Y. , Gou L., Peng M., Zheng J., and Chen L.. 2021. “Effects of Soluble Fiber Supplementation on Glycemic Control in Adults With Type 2 Diabetes Mellitus: A Systematic Review and Meta‐Analysis of Randomized Controlled Trials.” Clinical Nutrition 40, no. 4: 1800–1810. 10.1016/j.clnu.2020.10.032. [DOI] [PubMed] [Google Scholar]
- Xu, C. , and Doi S. A. R.. 2018. “The Robust Error Meta‐Regression Method for Dose‐Response Meta‐Analysis.” International Journal of Evidence‐Based Healthcare 16, no. 3: 138–144. 10.1097/xeb.0000000000000132. [DOI] [PubMed] [Google Scholar]
- Yoshida, Y. , Tamura Y., Kouzaki K., and Nakazato K.. 2023. “Dietary Apple Polyphenols Enhance Mitochondrial Turnover and Respiratory Chain Enzymes.” Experimental Physiology 108, no. 10: 1295–1307. 10.1113/EP091154. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supplementary Table 1. Search keywords and search strategy.
Supplementary Table 2. Query boxes of each database based on the Title & Abstract (search date: August 1, 2026).
Data Availability Statement
All data used in this meta‐analysis were extracted from published studies. The datasets supporting the findings of this study are available from the original sources, which are cited in the manuscript. Additional information can be provided by the corresponding author upon reasonable request.
