Supplements for Weight Loss on a GLP-1: 16 Supplements Evidence-Graded Against PubMed

Sixteen popular weight-loss supplements graded A through D against verified PubMed primary sources. Only three reach grade A or B for weight loss with credible RCT evidence: berberine (Asbaghi 2020 meta −2.07 kg)^[1], MCT oil as a long-chain triglyceride replacement (Mumme 2015 meta −0.51 kg vs LCT)^[4], and green tea catechins (Hursel 2009 meta −1.31 kg)^[5]. Glucomannan^[9], psyllium^[10], CLA^[8], and apple cider vinegar reach grade B with caveats. The remaining 9 supplements have grade C or D evidence — either animal-only data, a single underpowered trial, no human weight-loss data at all, or evidence so methodologically weak that it disappears under quality filters. Even the best supplements produce roughly 1-5% of GLP-1 magnitude. Lemon balm has 2,700 monthly searches and zero human weight-loss RCTs. Marketing volume does not equal evidence. Here is the verified evidence map.

The grading scale

• Grade A: Multiple RCTs converge on a consistent effect; meta-analysis with statistically significant magnitude; mechanism well-characterized.
• Grade B: Single decent meta-analysis or multiple RCTs with directional signal; modest magnitude; some methodological caveats.
• Grade C: Animal or pilot human data only; single underpowered RCT; or evidence that disappears under quality filters.
• Grade D: No human RCT evidence for weight loss specifically; folklore or marketing claims only.

Grade A: Berberine

We covered berberine in detail in our dedicated berberine vs GLP-1 article. The headline: Asbaghi 2020^[1] meta-analyzed 12 RCTs and reported a mean weight reduction of −2.07 kg (95% CI −3.09 to −1.05, p<0.001) and BMI reduction of −0.47 kg/m² over a median 12-week follow-up. Yin 2008^[2] anchors the diabetes side. The Liu 2010 PK study^[3] showed approximately 0.36% oral bioavailability in rats — the bottleneck that limits any “nature's Ozempic” magnitude argument.

Grade: A. Real, replicable, well-characterized mechanism, but ~5-8% of GLP-1 magnitude. Reasonable adjunct; not a substitute.

Grade B: MCT oil (as a fat replacement)

Mumme and Stonehouse 2015^[4] meta-analyzed 13 RCTs (n=749) of medium-chain triglycerides as a replacement for long-chain triglycerides in the diet. Pooled effect: −0.51 kg (95% CI −0.80 to −0.23, p<0.001) for body weight; −1.46 cm waist; −0.79 cm hip. Effect was consistent across trials but modest in absolute magnitude.

Critical caveat: MCT oil works as an LCT replacement, not as an additive caloric source. Bulletproof coffee (added MCT oil + butter on top of normal eating) is not the trial design. For patients to capture the modest benefit, they need to substitute MCT for an equal-calorie amount of LCT (typically olive oil or other long-chain fat).

Grade: B. Real but modest, ~1-2% of GLP-1 magnitude. Cost is not negligible (~$15-30/month for therapeutic doses).

Grade B: Green tea catechins / EGCG

Hursel 2009^[5] meta-analyzed 11 RCTs of green tea catechin preparations (typically standardized to ~270 mg EGCG/day with caffeine). Pooled effect: −1.31 kg body weight, p<0.001. The effect was modulated by ethnicity (Asian populations showed larger responses) and by habitual caffeine intake (regular caffeine consumers showed smaller effects due to catechin-caffeine interaction).

Decaffeinated green tea preparations show minimal benefit; the catechin + caffeine combination is what drives the effect. Hepatotoxicity has been rarely reported with high-dose green tea extract supplements (not with green tea as a beverage); USP and Cochrane note this as a low- but-real signal warranting label warnings.

Grade: B. Real but modest, ~3-5% of GLP-1 magnitude. Decaffeinated forms minimal benefit.

Grade B: Glucomannan and psyllium fiber

Glucomannan (konjac fiber): Sood 2008^[9] meta-analyzed 14 RCTs (n=531). Pooled body weight effect: −0.79 kg (95% CI −1.53 to −0.05). Also reduced total cholesterol by 19.28 mg/dL and triglycerides by 11.08 mg/dL. Mechanism: viscous fiber gels in the stomach, expanding gastric volume and delaying emptying — satiety, not metabolism.

Psyllium husk (Plantago ovata): Pal 2011^[10] randomized 66 overweight adults to psyllium plus a healthy diet over 12 weeks. Modest body composition and lipid improvements. Multiple subsequent meta-analyses confirm a similar profile: ~0.8-1.5 kg weight effect and meaningful LDL/triglyceride improvements.

Both fibers carry a real choking risk if not taken with adequate water. Both are inexpensive, safe in renal disease, and support the satiety side of GI tolerance on a GLP-1.

Grade: B. Real but modest, ~2-3% of GLP-1 magnitude. Inexpensive and safe; reasonable adjuncts.

Grade B: Conjugated linoleic acid (CLA)

Whigham 2007^[8] meta-analyzed 18 RCTs of CLA supplementation. At 3.2 g/day (the typical effective dose), fat mass loss was approximately −0.09 kg per week versus placebo, plateauing after about 6 months. That works out to roughly −4.7 kg of fat over 6 months at the high dose — not trivial, but with meaningful caveats: GI upset (diarrhea, fatty stools), insulin resistance reported in some diabetic subgroups, and cost ($30-50/month) that erodes cost-effectiveness.

Grade: B. Real but modest, ~3-5% of GLP-1 magnitude. Side effects and cost limit value.

Grade B: Apple cider vinegar (with caveats)

Apple cider vinegar is heavily marketed for weight loss. The published evidence is thin. The Khezri 2018 trial in the Journal of Functional Foods (not PubMed- indexed at search time) reported an additional ~1.2 kg weight loss vs diet alone in 39 patients on a calorie- restricted diet. Launholt 2020^[14] in European Journal of Nutrition systematically reviewed the ACV literature and concluded that evidence for weight or metabolic effects is insufficient due to methodological limitations across the trial base.

A separate 2024 BMJ Nutrition Prevention Health trial in Lebanese adolescents (Abou-Khalil 2024) initially reported large effects but was retracted in September 2025 due to improbable data characteristics. Treat any weight-loss ACV claim sourced to that paper as void.

Side effects: dental erosion with prolonged daily exposure; esophageal irritation if undiluted. Cost: ~$1-3/month (cheap vinegar from a grocery store).

Grade: B with caveats. Effect is small, diet-context dependent, and the literature is methodologically weak. Cheap and harmless in moderation, but unlikely to meaningfully move the needle.

Grade C: Ashwagandha (Withania somnifera)

Choudhary 2017^[12] randomized 52 chronically stressed adults to ashwagandha or placebo for 8 weeks. The trial reported significant improvements in stress, food cravings, and a directional improvement in body weight, but the exact weight effect size was not disclosed in the available abstract. The weight-loss arm of the ashwagandha evidence base is essentially this one underpowered trial in a stressed-adult population.

Cortisol-reduction evidence (Salve 2019, Lopresti 2019, covered in our stress and cortisol article) is somewhat stronger, but cortisol reduction is not weight loss.

Grade: C. Plausible adjunct for stress eating; not a primary weight intervention. Watch for rare hepatotoxicity case reports.

Grade D / oppositional: Creatine monohydrate

Creatine is the most-misunderstood supplement on the list. It is one of the most rigorously-studied performance supplements with strong evidence for muscle preservation and strength gains during resistance training. But for weight loss, creatine works in the opposite direction: initial water retention from muscle creatine loading typically increases body weight by 0.5-2 kg in the first weeks. Long-term, creatine supports lean mass preservation in caloric deficit (Forbes 2019 meta-analysis: lean mass +0.68 kg, especially when paired with resistance training).

For a GLP-1 patient who is already losing fat and trying to preserve lean mass, creatine 5 g/day plus resistance training is a defensible adjunct — it tilts the body composition outcome more favorably without preventing fat loss. But it does not cause weight loss and will increase the scale number short-term.

Grade: D for weight loss; A for muscle preservation when used with resistance training. See our exercise pairing article.

Grade C: Magnesium

Magnesium is metabolically important and many adults are marginally deficient. Recent meta-analyses (Askari 2021 and others) of magnesium supplementation and body weight report modest BMI reductions (~−0.21 kg/m²) without significant body weight effects overall. Subgroup analyses in obese patients show waist circumference reductions of ~2 cm. Magnesium is useful for muscle cramps, sleep, and insulin sensitivity in deficient patients; it is not a weight-loss agent.

Grade: C for weight loss; B for related metabolic support in deficient adults.

Grade D: Collagen peptides

The Proksch 2014 RCT^[13] randomized 69 women to oral collagen peptides or placebo for 8 weeks. The trial showed significant improvement in skin elasticity (the primary endpoint), but did not measure or report weight loss. Collagen peptides are protein, so they contribute to protein satiety like any other protein source, but there is no evidence that the collagen-specific composition produces weight loss above and beyond what an equivalent protein dose would.

Useful for skin elasticity in midlife women; useless for weight loss directly. Relevant for patients losing weight rapidly on a GLP-1 who are concerned about loose skin (see our loose skin article).

Grade: D for weight loss; B for skin elasticity in healthy mid-life women.

Grade C: Cinnamon

Cinnamon has a weak, statistically detectable effect on fasting glucose in T2D patients (Allen 2013 Annals of Family Medicine meta-analysis: ~−24.6 mg/dL fasting glucose, no significant A1c effect). The Allen meta did not report weight as a primary outcome and the broader cinnamon literature does not support a meaningful weight effect. The TikTok “cinnamon coffee for weight loss” trend is not evidence-based.

Grade: C. Modest glucose effect, no meaningful weight effect.

Grade D: Garcinia cambogia (HCA)

Garcinia cambogia (hydroxycitric acid) is one of the most heavily marketed weight-loss supplements. Onakpoya 2011^[6] meta-analyzed 12 RCTs and reported a pooled weight effect of −0.88 kg (95% CI −1.75 to 0.00) — statistically borderline. When the analysis was restricted to high-quality RCTs only, the effect disappeared entirely. The signal in the meta is driven by low-quality, short-duration, small-sample trials.

Safety: rare case reports of hepatotoxicity (FDA Consumer Update 2009, multiple subsequent case reports). The magnitude is too small to justify any risk.

Grade: D. Minimal effect that disappears with quality filtering. Marketing exceeds evidence substantially.

Grade C: Chromium picolinate

Onakpoya 2013^[7] meta-analyzed 20 RCTs of chromium picolinate. The pooled body weight effect was statistically significant but the authors explicitly noted that the clinical magnitude was unclear and the effect was driven by trials of variable quality. Chromium is metabolically active in patients with documented chromium deficiency (rare); for the general population it does not produce meaningful weight loss.

Grade: C. Statistically detectable but clinically marginal.

Grade D: Lemon balm (Melissa officinalis)

Lemon balm has 2,700 monthly searches in the US for weight loss. The Heshmati 2020^[11] meta-analysis of 7 RCTs on lemon balm and cardiometabolic outcomes explicitly did not measure body weight. Lipids, blood pressure, and glucose were unchanged. The remaining lemon balm literature is on anxiety and stress, where it shows modest benefit. There is no human RCT evidence for lemon balm as a weight-loss intervention.

Grade: D. Marketing volume vastly exceeds evidence. Useful for anxiety; not for weight.

Grade D: L-Lysine

L-lysine has 2,300+ monthly searches for weight loss. Our verification subagent searched PubMed extensively and found no human RCT evidence for L-lysine causing weight loss. The mechanism stories (carnitine biosynthesis, ketone metabolism) are speculative. More than 94% of US adults already meet the WHO/FAO lysine requirement from diet alone, so supplementation is unlikely to fill any meaningful nutritional gap.

Grade: D. UNVERIFIED for weight loss in humans.

The summary table

Comparator anchors: GLP-1 magnitudes

• Semaglutide 2.4 mg (STEP-1): −14.9% body weight at 68 weeks^[15]
• Tirzepatide 15 mg (SURMOUNT-1): −20.9% at 72 weeks^[16]

For a 100 kg starting weight, that's −15 to −21 kg. Even the highest-grade supplement (berberine at −2.07 kg) is roughly 1/7 to 1/10 the magnitude. Patients seeking 10%+ weight loss are not going to get there on supplements alone.

Regulatory context

Dietary supplements in the US cannot legally make disease treatment claims (DSHEA 1994). Weight loss is not a disease per se but weight-loss claims still attract FTC scrutiny when unsupported. FDA does not pre-approve supplement efficacy or label accuracy. Independent testing programs (USP, NSF, ConsumerLab) repeatedly find purity and potency variation across supplement brands, especially in weight-loss product categories.

The NIH Office of Dietary Supplements maintains a weight-loss supplement factsheet that is more conservative than published meta-analyses and emphasizes clinical relevance over statistical significance. Their bottom line aligns closely with this article: the evidence base for weight-loss supplements is thin, and patients seeking meaningful weight loss should focus on lifestyle changes and FDA-approved pharmacotherapy.

The honest patient framing

• If you want significant weight loss (≥10%): supplements are not going to deliver. GLP-1s (semaglutide, tirzepatide, orforglipron), Qsymia, or bariatric surgery are the evidence-based options.
• If you're already on a GLP-1 and want a metabolic-support adjunct: berberine (if not on a CYP3A4-sensitive medication), green tea catechins, and fiber (glucomannan or psyllium) have the strongest supplement evidence.
• If you're trying to preserve lean mass on a GLP-1: creatine 5 g/day plus resistance training is the best-evidenced supplement intervention. See our exercise pairing article.
• If you're on a GLP-1 and worried about loose skin: oral collagen peptides have evidence for skin elasticity (Proksch 2014) but not for weight loss. See our loose skin article.
• Skip: garcinia cambogia, chromium picolinate, lemon balm for weight loss, L-lysine for weight loss, and any supplement marketed as “nature's Ozempic.”

Bottom line

• 16 popular weight-loss supplements graded against PubMed primary sources.
• Only berberine reaches grade A. Only six others reach grade B. Nine are grade C or D.
• Even the best supplements produce roughly 1-5% of GLP-1 magnitude.
• Marketing volume does not equal evidence: lemon balm and L-lysine have thousands of monthly searches and zero human weight-loss RCTs.
• Creatine is oppositional: it increases scale weight short-term but supports lean mass preservation when paired with resistance training.
• For meaningful weight loss, FDA-approved pharmacotherapy (GLP-1s or Qsymia) and lifestyle change are the evidence-based options.

Related research and tools

• Berberine vs GLP-1: dedicated deep-dive
• Exercise pairing on a GLP-1 — for the creatine + resistance training context
• Loose skin after rapid GLP-1 weight loss — for the collagen peptide context
• Stress, cortisol, and food noise on a GLP-1 — for the ashwagandha context
• Metformin and non-GLP-1 diabetes drugs for weight loss — the prescription comparators
• GLP-1 protein calculator — for the actual evidence-based dietary intervention

Important disclaimer. This article is educational and does not constitute medical advice. Dietary supplements are not FDA-approved for weight loss and may interact with prescription medications. Patients on statins, anticoagulants, antidepressants, thyroid medication, or any GLP-1 receptor agonist should discuss supplement use with their prescribing clinician before starting. Berberine inhibits CYP3A4 and warrants caution with statins (see our berberine article). Green tea extract has rare hepatotoxicity case reports at high doses. Garcinia cambogia has been associated with rare liver injury. Every primary source cited here was independently verified against PubMed on 2026-04-08. Items the verification subagent could not confirm against primary sources (specifically L-lysine and lemon balm for weight loss) are explicitly flagged as UNVERIFIED rather than paraphrased.