Citrus Flavonoids With Metformin: One Small Trial

Can you add a citrus polyphenol supplement to metformin and get better prediabetes results? One small randomized trial offers a qualified yes: participants taking a blend of eriocitrin, hesperidin, and naringin had about 5% lower 2-hour oral glucose-tolerance-test (OGTT) glucose and about 12% lower TNF-α than when they took placebo. That is a signal, not a treatment standard.

The difference matters most if you are near the diagnostic boundary for type 2 diabetes. A 5% change in a 2-hour glucose value could amount to several milligrams per deciliter, although the study reported a relative attenuation rather than an absolute mg/dL difference. Whether that change prevents diabetes, improves HbA1c over years, or adds anything meaningful to diet and exercise remains unknown.

What the metformin-plus-flavonoid trial found

Cesar et al.'s 2025 trial enrolled 71 adults with prediabetes. In this double-blind, placebo-controlled crossover study, participants continued metformin at 1,000 mg per day and received either placebo or 250 mg per day of a citrus flavonoid blend containing eriocitrin, hesperidin, and naringin. Each treatment period lasted 12 weeks, separated by a two-week washout; 35 participants began in one sequence and 36 in the other.

During the flavonoid period, 2-hour OGTT glucose fell by approximately 5% relative to placebo, and TNF-α, an inflammatory cytokine linked to insulin resistance, fell by approximately 12%. The paper also refers to a companion study by the same group that reported a 17% increase in circulating GLP-1 after 12 weeks with the same supplement in people with prediabetes and diabetes.

Those outcomes point in a favorable direction, but they are surrogate measures. A single OGTT value is not the same as lower diabetes incidence, fewer cardiovascular events, or a durable reduction in HbA1c. The publication excerpt available for appraisal does not provide the relevant p-values, confidence intervals, or between-period effect estimates. That prevents readers from judging the precision of the glucose and TNF-α findings or deciding whether chance could plausibly explain them.

Why the combination makes biological sense

Eriocitrin, hesperidin, and naringin mostly reach the colon intact, where gut microbes can remove their sugar groups and produce the aglycones eriodictyol, hesperetin, and naringenin. Reviews of polyphenol research describe several plausible actions: AMPK activation, DPP-4 inhibition, longer endogenous GLP-1 activity, and lower inflammatory signaling such as TNF-α and IL-6. A 2022 review of fruit and vegetable polyphenols summarizes that framework.

Metformin also affects hepatic AMPK signaling and lowers hepatic glucose production. The overlap creates a reasonable hypothesis for an added glycemic effect. It does not prove synergy. A trial must compare the combination with metformin alone, as this one did, but it also needs replication before a mechanistic story earns clinical confidence. Cell studies and animal pharmacology can explain a result after the fact; they cannot substitute for a second human trial.

The wider polyphenol evidence supports plausibility, not this formula

Other randomized trials show that people with prediabetes can respond metabolically to polyphenol-rich foods or botanicals. In a 12-week, double-blind, placebo-controlled trial, moringa leaf powder at 2,400 mg per day produced more favorable changes in fasting glucose and HbA1c than placebo among adults with prediabetes. A randomized crossover trial of red raspberries, with or without fructo-oligosaccharides, found improved insulin sensitivity and changes in gut microbiota in adults with prediabetes, while HbA1c effects were limited.

A 2025 strawberry crossover trial also reported improved insulin resistance and related cardiometabolic markers in adults with prediabetes. Separately, a 12-week trial in 120 Asian Indian adults with prediabetes tested 60 g of pistachios per day—30 g before breakfast and 30 g before dinner—and reported improvements in cardiometabolic risk factors.

These studies make the citrus result less biologically surprising. They do not validate the exact 250 mg blend, its individual ingredients, or its use with metformin. Different foods contain different polyphenols, doses, fibers, fats, and calories. They also use different endpoints and populations. "Polyphenols help" is too broad to guide a supplement purchase.

Why one positive trial should not change your regimen

Every direct human glycemic finding for this particular citrus blend comes from one team. That does not make the result wrong. It does mean an independent group has not yet tested whether the effect repeats in a new population, with a disclosed formulation, full statistical reporting, and a design built to examine safety.

The crossover design helps a small study because each participant can serve as their own comparator. The two-week washout creates a legitimate question, however. These flavonoids depend partly on colonic microbial metabolism, and dietary interventions can change the microbiome. The red-raspberry trial documented microbiota changes, but it did not show that this citrus blend causes changes that persist beyond two weeks. The concern is theoretical, not a demonstrated flaw in the citrus trial. Still, a future study should use a longer washout or a parallel-group design.

The formulation also leaves practical gaps. The study reports 250 mg per day of the three-compound blend but does not disclose the eriocitrin-to-hesperidin-to-naringin ratio. No human dose-ranging data establish whether 250 mg is near a minimum effective dose, a plateau, or an unnecessarily high dose. You cannot reproduce a study precisely when the active blend is only partly specified.

Safety needs human pharmacokinetic data

Citrus aglycones such as naringenin and hesperetin have raised a preclinical concern: they may inhibit OCT2 and MATE transporters involved in renal metformin elimination. If that happens to a clinically meaningful degree in humans, metformin exposure could rise. The current trial does not establish such an interaction, and the available evidence does not quantify it in people.

That uncertainty matters most for people with impaired kidney function, substantial gastrointestinal intolerance, or several medications. The trial excluded people with chronic gastrointestinal disorders, so it does not answer how well the supplement works—or how it feels—in people with IBS, IBD, or persistent metformin-related GI symptoms. Anyone considering the combination should discuss it with the clinician who prescribes their metformin, especially if kidney function or other medications complicate the picture.

How this compares with evidence-backed add-ons

The bar for a drug add-on is much higher. A systematic review of randomized trials of SGLT-2 inhibitors added to metformin in type 2 diabetes found HbA1c reductions of roughly 0.6 to 0.9 percentage points versus placebo across multiple independent trials and populations. That evidence base involves a different disease stage and endpoint, so it is not a head-to-head contest with citrus flavonoids. It does show the difference between an early nutraceutical signal and an intervention with replicated clinical evidence.

For prediabetes itself, lifestyle intervention has deeper support. A systematic review and meta-analysis of aerobic exercise trials included 10 randomized trials and 815 people with prediabetes and found improvements in blood glucose, HbA1c, and BMI. Exercise lacks the romance of a capsule, but it has more than one small study behind it.

Verdict: the citrus blend is interesting enough to watch, but not established enough to treat as a proven metformin enhancer.

The current evidence supports a low-to-moderate confidence judgment. If an independent randomized trial confirms the OGTT and GLP-1 findings with complete statistical reporting and a longer washout, confidence should rise. Human pharmacokinetic data showing no meaningful change in metformin exposure would answer the most important safety question. Until then, the result is one encouraging trial—not a protocol.