As a complex and chronic medical condition, obesity is a worldwide health problem (Thomas et al., 2014) and it is a cluster of complications such as impaired glucose tolerance, dyslipidemia, hypertension, and systemic inflammation (Stoner and Cornwall, 2014). It is also a leading cause of extensive morbidity and mortality with a high economic burden in both developing and developed nations (Lehnert et al., 2013, Ng et al., 2014). Obesity is one of the direct sources for progression or occurrence of various diseases such as cardiovascular diseases (Mandviwala et al., 2016), insulin resistance and diabetes (Genser et al., 2016), non-alcoholic fatty liver disease (Li et al., 2016), gallstone disease and pancreatitis (Bonfrate et al., 2014), esophageal reflux (Khan et al., 2016), inflammatory bowel disease (Harper and Zisman, 2016), chronic kidney disease (Briffa et al., 2013), poly ovary syndrome (Orio et al., 2016), neurological diseases (Martin-Jiménez et al., 2017), and different types of cancer (Donohoe et al., 2017). It has been reported that the prevalence of overweight and obesity has been growing from 29% to 38% in the past three decades (Ng et al., 2014). Recent reports from World Health Organization in 2016 have suggested that 1.9 billion people are overweight, of whom 650 million are considered to be obese (World Health Organization, October 2017).
Different weight management strategies have been used throughout the years, including dietary regimen approaches or dietary constituents with anti-obesity potentials (Tuomilehto et al., 2001). However, some of the traditional recommendations have been proved to be impractical and disappointing due to poor attendance and adherence rates (Butryn et al., 2011, Huisman et al., 2010, Zuckoff, 2012). Thus, novel strategies are desperately needed to be suggested as more effective ways to get rid of excessive fat (Soeliman and Azadbakht, 2014). A variety of weight loss supplements sold under the title of “slimming aids” is already available although the outcome of each remains uncertain (Derosa and Maffioli, 2012, Mousavi et al., 2018, Onakpoya et al., 2011). It has been demonstrated that dietary intake of a widespread group of plant polyphenols known as flavonoids can exert anti-obesity effects; however, enough evidence is not available in this regard (Bertoia et al., 2016). Some trials suggest that a potential effect of polyphenols can reduce body weight by increasing energy expenditure (Barth et al., 2012, Dallas et al., 2014, Most et al., 2014) even though other findings have shown null or contrasting results (Bell et al., 2011, Janssens et al., 2015).
Hesperidin is a flavonone glycoside that along with narirutin is a subgroup of flavonoids that mainly exist in the solid parts of citrus fruits and the membranes separating the pulp segments; this explains why the concentration of these flavonones is higher in whole-fruit juices (Roowi et al., 2009). There is evidence on the cardio-protective effects and anti-inflammatory properties of hesperidin (Amiot et al., 2016, Lorzadeh et al., 2019, Mulvihill et al., 2016). However, some of these results are inconsistent with recent meta-analyses suggesting hesperidin supplementation might not have any impact on lipid profile, blood pressure, and blood glucose control (Mohammadi et al., 2019, Shams-Rad et al., 2020). Although there is still no strong evidence on the association between flavanones and weight loss, some studies have investigated hesperidin as a supplement for obesity management and reported that it may be useful for the prevention or treatment of obesity (Ohara et al., 2016). However, several other studies have not supported these claims (Demonty et al., 2010, Ribeiro et al., 2017, Simpson et al., 2016) . Therefore, this study aims to summarize the available data of randomized controlled clinical trials (RCTs) investigating the effect of hesperidin supplementation on anthropometric parameters in adults.
Materials and Methods
The preferred reporting items of systematic reviews and meta-analysis (PRISMA) statement was followed to design perform and report this systematic review and meta-analysis (Shamseer et al., 2015).
Data sources and search strategy: Medline/Pubmed, Scopus, Web of Science, and Google Scholar were systematically searched from the earliest available online indexing through February 2022. The search was limited to human studies, additionally, no language restriction was performed. Two investigators (Ramezani-Jolfaie N and Mohammadi M) independently assessed the relevancy of studies by their title and abstracts as well as full text if needed in the next step. Additionally, the references of selected articles were examined manually for any missing related studies. Three groups of medical subject heading terms (MeSH) and non-MeSH keywords were used in constructing the database search as follows: group 1: “hesperidin”, “hesperitin”, “citrus flavonoid”, “orange juice”; group 2: “intervention”, “trial”, “randomized”, “random”, “randomly”, “placebo”, “assignment”, “clinical trial”, “RCT”, “cross-over”, “parallel”, “body weight”, weight, “body mass index”, BMI, “waist circumference”, WC, “waist-hip ratio”, WHR, “hip circumference”, HC, “fat free mass”, FFM, “fat mass”, FM, “lean body mass”, LBM; group 3: “mouse”, “mice”, “rats”, “in vitro”, “pig”, “rabbit”, “rooster”, “cell”, “cow” that combined by utilizing the “NOT” Boolean operator.
Selection criteria: The Population, Intervention, Comparison, Outcome, and Study types (PICOS) are provided in Table 1. The original RCTs were considered for inclusion if they had supplemented hesperidin in human adults. Studies were excluded if they had a short duration of intervention (lower than 2 weeks), were conducted among children/adolescents below 18 years of age, and if they had no control/placebo comparison group or outcomes of interest or reported duplicate data. Studies with interventions containing the other components in addition to the hesperidin were also excluded.
Data extraction: Study details were extracted and recorded independently by two authors (Ramezani-Jolfaie N and Mohammadi M) and in case of any disagreement, a third party (Salehi-Abargouei A) was consulted to reach mutual consensus. The following information of the included RCTs was collected: study design (crossover or parallel), ethnic or country, the last name of each author, year of article publication, subject baseline characteristics (sample size, gender, age, and overall health status), intervention duration, the use of run-in or washout periods, dose of hesperidin intake (mg/day), type of intervention used in the control groups, and the number of participants who completed the follow-up period.
Risk of bias assessment: Risk of bias assessment was conducted against the following key criteria according to the recommendation of the Cochrane Collaboration (Higgins and Green, 2011): random sequence generation; allocation concealment; blinding of participants, personnel and assessors; incomplete outcome data; selective outcome reporting and other sources of bias. Determination of bias level was either low (proper methods taken to reduce bias), high (improper methods resulting in bias), or unclear (either a lack of sufficient information or uncertainty over a potential bias) risk of bias. Six domains were presented as the (‘key domains’) and used to decide whether each RCT was low risk (low for all key domains), high risk (high for one or more key domains), and unclear risk (unclear for at most one or more key domains). Any discrepancies were resolved by consulting with a third author if necessary (ASA).
Data analysis: The mean change values with their respective standard deviations (SDs) between baseline and end of the study in both treatment and control groups were extracted. If change values were not reported by the studies, the baseline and final mean values with their respective SDs were used to calculate the mean±SD of changes in outcomes by a correlation coefficient of 0.5. Further analyses using r = 0.1 and 0.9 were also performed to check the sensitivity of the findings to the selected correlation coefficient. The random-effects model was used to examine the weighted mean differences (WMDs) and 95% confidence intervals (CIs). To assess the between-study heterogeneity, I2 statistics and Chi-square were incorporated which considered as significant and high by P-values < 0.05 for Chi-square and I2 values of more than 50%, respectively. Subgroups analysis was performed according to intervention duration (<8 weeks or >=8 weeks), dosage of hesperidin (<= 500 mg/day or > 500 mg/day), and health status of individuals (cardio-metabolic disorders or healthy) to explore the potential sources of heterogeneity. To test the robustness of the meta-analysis results, sensitivity analysis was also conducted by removing one trial at a time and recalculating the overall effects with the remaining studies. Moreover, the visual inspection of the funnel plots represented the possibility of publication bias (Egger et al., 2008). All the analyses were done by STATA software version 13.0 (StataCorp, Texas, USA) for which p-values less than 0.05 were considered statistically significant.
Results
Search results
In general, out of 6118 published articles that initially identified through systematic data search, 2180 were detected as duplicates and 3913 were excluded after title and abstract screening for not meeting the inclusion criteria. A total of 25 articles remained for full text evaluation from which 16 articles were further excluded for the following reasons: (i) did not report the outcomes of interest (n=6) (Cheraghpour et al., 2019, Kean et al., 2015, Martínez-Noguera et al., 2020, Milenkovic et al., 2011, Salden et al., 2016), (ii) had interventions containing other components in addition to hesperidin (n=6) (Rangel-Huerta et al., 2015, Valls et al., 2021a, Valls et al., 2021b, Yari et al., 2021c, Yoshitomi et al., 2021), (iii) had a short duration of intervention (<2 weeks, n=2) (Lamport et al., 2016, Schär et al., 2015), (iv) had no suitable control group (n=1) (Miwa et al., 2004), (v) reported duplicate data (n=1) (Homayouni et al., 2018). Nine studies were finally included in the present systematic review and meta-analysis (Demonty et al., 2010, Eghtesadi et al., 2016, Haidari et al., 2015, Homayouni et al., 2017, Ohara et al., 2016, Rizza et al., 2011, Yari et al., 2021a, Yari et al., 2021b, Yari et al., 2019). Seven out of nine articles have reported data on body weight (Demonty et al., 2010, Eghtesadi et al., 2016, Haidari et al., 2015, Homayouni et al., 2017, Ohara et al., 2016, Yari et al., 2021b, Yari et al., 2019), eight on BMI (Demonty et al., 2010, Eghtesadi et al., 2016, Haidari et al., 2015, Homayouni et al., 2017, Ohara et al., 2016, Rizza et al., 2011, Yari et al., 2021a, Yari et al., 2021b), five on waist circumference (WC) (Haidari et al., 2015, Ohara et al., 2016, Rizza et al., 2011, Yari et al., 2021a, Yari et al., 2019), two on hip circumference (HC) (Haidari et al., 2015, Ohara et al., 2016), and two on waist to hip ratio (WHR) (Haidari et al., 2015, Yari et al., 2021a) (Figure 1).
Study characteristics
The basic characteristics of RCTs are presented in Table 2. Nine randomized trials published from 2010 to 2020 were included, 8 of which had a parallel design and one was a cross-over trial (Rizza et al., 2011). Nine studies included 493 participants and sample sizes ranged from 24 to 124 participants of both sexes, aged between 18 to 75 years. The treatment duration lasted for 3 to 12 weeks and the dosage of hesperidin oral administration varied from 500 mg/day to 1000 mg/day. The majority of studies have been conducted in Iran, but one was based in Netherland (Demonty et al., 2010), one in Italy (Rizza et al., 2011), and another one in Japan (Ohara et al., 2016). The participants were either healthy with moderate obesity (Ohara et al., 2016) and moderate hypercholesterolemia (Demonty et al., 2010) or patients with a medical condition such as metabolic syndrome, myocardial infarction, diabetes, and non-alcoholic fatty liver diseases.
Risk of bias assessment: The risk of bias assessment of each RCT is provided in Table 3. Four studies out of nine used an adequate random sequence generator and had a low risk of bias for this domain (Haidari et al., 2015, Homayouni et al., 2017, Yari et al., 2021a, Yari et al., 2019), whereas the remaining five studies had an unclear risk of bias since no detailed method was suggested for randomization (Demonty et al., 2010, Eghtesadi et al., 2016, Ohara et al., 2016, Rizza et al., 2011, Yari et al., 2021b). Only one study by Homayouni et al. (Homayouni et al., 2017) applied adequate allocation concealment and other articles did not report clear data on allocation concealment, hence considered as unclear risk of bias. Two studies were categorized as high risk of bias for blinding of the participants and personnel (Yari et al., 2021a, Yari et al., 2021b); however, by providing enough information about blinding in the remaining studies, they were labeled as low risk of bias. There was no sufficient report on blinding of outcome assessment in any of the included studies. The risk of bias from incomplete outcome data was assessed as low in the majority of the articles except for two (Haidari et al., 2015, Rizza et al., 2011). Regarding selective outcome reporting, all the studies were judged as low risk of bias. As a result, in overall risk of bias, seven of the included studies were assessed as “unclear”, since each study had an unclear risk of bias for at least one of the six domains. Two remaining trials were regarded as “high” risk of bias due to having at least one high-risk domain.
Systematic review
Effect of hesperidin supplementation on HC and WHR: Two studies (n = 104 participants) on HC (Haidari et al., 2015, Ohara et al., 2016) and two studies (n = 118 participants) on WHR (Haidari et al., 2015, Yari et al., 2021a) provided no evidence for the effectiveness of hesperidin in reducing HC and WHR values. Haidari et al. (Haidari et al., 2015) reported no significant differences in HC and WHR at baseline and at the end of the study between hesperidin and placebo groups (P > 0.05). Ohara et al. (Ohara et al., 2016) also showed no significant differences in reducing HC between subjects receiving placebo and those who ingested glucosyl hesperidin with or without caffeine. In a study by Yari et al. (Yari et al., 2021a), hesperidin supplementation also resulted no significant changes in WHR compared to control groups.
Meta-analysis
Effect of hesperidin supplementation on body weight: As provided in Table 4, the pooled estimated effect size of seven studies with 426 participants (Demonty et al., 2010, Eghtesadi et al., 2016, Haidari et al., 2015, Homayouni et al., 2017, Ohara et al., 2016, Yari et al., 2021b, Yari et al., 2019) showed no significant changes in body weight after hesperidin consumption compared to control groups (WMD=0.01 kg, 95% CI: -0.22, 0.24, P=0.918; Figure 2). There was no significant between-study heterogeneity (Q statistics=3.54, P=0.739, I2=0%). We also performed some subgroup analyses to identify the possible different effects of hesperidin supplementation caused by duration and dosage of treatment and health status of the participants; however, no significant changes in weight status were observed in any of the subgroups.
Effect of hesperidin supplementation on BMI: Eight studies with 444 participants were assessed for effects of hesperidin on BMI (Demonty et al., 2010, Eghtesadi et al., 2016, Haidari et al., 2015, Homayouni et al., 2017, Ohara et al., 2016, Rizza et al., 2011, Yari et al., 2021a, Yari et al., 2021b). Hesperidin supplementation was found to have no significant effect on BMI in comparison with control groups (WMD=-0.02 kg/m2, 95% CI: -0.16, 0.13, P=0.831; Figure 3). There was moderate between-study heterogeneity, but it was not significant (Q statistics=9.58, P=0.214, I2 =26.9%). No statistical difference was observed in subgroups according to duration and dosage of treatment and health status of the participants (Table 4).
Effect of hesperidin supplementation on WC: The overall effects of 5 studies including a total of 220 participants (Haidari et al., 2015, Ohara et al., 2016, Rizza et al., 2011, Yari et al., 2021a, Yari et al., 2019) suggested no significant change in WC after hesperidin supplementation compared to control groups (WMD=-0.48 cm, 95% CI: -1.52, 0.55, P=0.362; Figure 4). A significant between-study heterogeneity was found (Q statistics=9.76, P=0.045, I2=59%). When the subgroup analysis was performed based on duration and dosage of treatment, heterogeneity was attenuated and non-significant in their categories; however, no significant change in WC was observed in any of the subgroups (Table 4).
Meta-regression
To examine the possible association of different effects of hesperidin with supplementation dose and study duration on body weight, the meta-regression analysis was performed, but no significant relationship was detected.
Publication bias and sensitivity analysis
The pooled effects of hesperidin intake on weight, BMI, and WC were not sensitive to any of the studies after omitting each out of analyses, suggesting the results were robust. Furthermore, correlation coefficients opted to examine the value changes in the meta-analyses revealed no indication regarding sensitivity.
No publication bias was perceived in the funnel plots and they proved to be symmetrical after considering Begg’s and Egger’s asymmetry tests: weight (Begg’s test, P=0.133; Egger’s test, P=0.027), BMI (Begg’s test, P=0.266; Egger’s test, P=0.062), WC (Begg’s test, P=0.086; Egger’s test, P=0.051).