Which supplements are actually worth taking? The short, evidence-based list

If you eat a reasonably varied diet, most of what is in a supplement aisle does nothing for you. The list with real randomised-trial evidence is short: creatine, vitamin D if you are actually deficient, omega-3, protein powder used as convenience, and caffeine. Everything else is, at best, optional, and a lot of the best-sellers are marketing dressed up as science.

A $60bn aisle built on a loophole

The US supplements market was worth somewhere between $60bn and $78bn in 2025 depending on which research firm you ask, and roughly three-quarters of US adults take something daily. That is not a niche habit. It is a near-universal one, and it runs on a regulatory gap most buyers do not know about.

In the US, the FDA does not review or approve supplements for safety or effectiveness before they go on sale. Drugs need pre-market approval; supplements do not. The burden falls on the FDA to prove a product unsafe after it is already on shelves. So "it is on the shelf, therefore someone checked it" is simply false. Nobody checked it before you bought it.

That gap is the whole story. Thousands of products compete on packaging and claims rather than trials, because the law lets them.

Food first, and what "filling a gap" really means

For almost everyone, whole foods beat pills. Real food brings fibre, water, and a mix of nutrients that a capsule strips out. Supplements earn their place only when they fill a specific, identified gap, or when they do something food cannot conveniently do.

That framing matters because it flips the default. The question is not "what should I add?" It is "is there an actual deficiency or a measurable performance reason here?" If the answer is no, the honest move is to spend the money on better groceries.

Bottom line

Supplements are a top-up for a specific gap, not a substitute for eating well. If your diet is varied and adequate, most products in the aisle do nothing measurable for you.

The short evidence-backed list

Creatine monohydrate. This is the strongest case in the building. It reliably increases lean mass and strength alongside resistance training, with no serious adverse effects in trials running up to two years at standard doses of around 3 to 5 grams a day. A 2025 review found it helps older adults preserve muscle, and a 2024 review found early signals for memory and attention, though that cognitive evidence is not yet settled. The myths are wrong: it is not just for bodybuilders, it does not damage healthy kidneys in trials, and the early weight bump is water inside the muscle cell, not fat. Caveat: the cognitive benefits are promising, not proven.

Vitamin D, only if you are deficient. A pooled analysis of 7.9 million people found about 15.7% had seriously low vitamin D, with far higher rates in older adults and at higher latitudes. For those people, modest doses help. For people who are already replete, blanket high-dose use shows no benefit. The caveat is sharp: vitamin D is fat-soluble and stored in the body, so it can reach toxic levels. More is not better, and "everyone should take it" is wrong.

Omega-3 (EPA/DHA). The cardiovascular evidence is mixed and dose-dependent. Some large meta-analyses link it to fewer cardiovascular events, but broad low-dose population trials have been largely null. The clearest benefit shows up at higher or purified prescription doses, and in people with elevated triglycerides. Caveat: eating oily fish is the first-line move, and a low-dose capsule is not a heart-attack shield for the general population.

Protein powder, as convenience. A landmark meta-analysis showed protein supplementation produces a small, real boost to muscle and strength during resistance training, but the benefit plateaus once total intake reaches roughly 1.6 g/kg/day. Whole-food protein does the same job. Powder is a tool for hitting a daily target on a busy schedule, not a special muscle-building compound. Caveat below: it is also one of the products where contamination testing matters most.

Caffeine. One of the best-supported performance aids. Doses of roughly 1 to 7 mg/kg produce small but consistent gains in strength, muscular endurance, alertness, jumping and sprinting. Caveat: regular heavy intake blunts the acute effect, so the everyday triple espresso habit works against you on the day it counts.

The benefit of protein supplementation plateaus once total intake reaches roughly 1.6 grams per kilo of bodyweight per day.Morton et al. protein meta-analysis

Quality and contamination: the label may not match the bottle

Because nobody vets these products before sale, what is on the label is not guaranteed to be in the bottle. A review of adulteration in sports supplements found that analytical studies have flagged anywhere from roughly 15% to 50% of tested samples as containing anabolic agents or other prohibited substances, with pre-workout, thermogenic and muscle-building products the most susceptible categories.

Protein powder is a live example. A 2025 Clean Label Project report tested 165 protein products from 70 brands and found 47% exceeded California Proposition 65 thresholds for heavy metals like lead and cadmium, with organic and plant-based powders worse. Industry bodies disputed the methodology, but the point stands: batch-to-batch testing is the only thing that tells you what you are actually swallowing.

That is where third-party certification comes in. Informed Sport tests every single batch for banned substances, contaminants and label accuracy. NSF Certified for Sport screens for 100-plus contaminants and banned substances, verifies the label, and adds annual facility audits plus marketplace sampling. The US Anti-Doping Agency points to NSF Certified for Sport as the best way to cut your risk. A "natural" badge means nothing; "natural" is a marketing word, not a safety or quality standard. Independent batch testing is the only quality signal that means anything.

Mostly noise

The popular sellers whose marketing badly outruns their evidence:

Multivitamins for healthy adults. A 2024 analysis of more than 390,000 healthy US adults followed for over 20 years found no association between daily multivitamin use and lower risk of death, and no benefit for cancer, heart or cerebrovascular mortality. For healthy people, diet quality matters far more than a daily pill.
BCAAs and glutamine. If your overall protein intake is adequate, the effect on body composition is negligible. You are paying extra for amino acids you already get from food and protein powder.
Testosterone boosters. The label claims routinely outrun the evidence.
Greens powders. They strip out the water and fibre that make vegetables worth eating in the first place.
Collagen. Heavily marketed, with evidence that runs out well short of the claims.

Dose, safety, and the "natural means safe" trap

Two ideas to retire. First, "natural means safe" — it does not. Natural is not a safety standard, a purity standard, or a testing standard. Second, "more of a good thing is better." Several nutrients, especially fat-soluble vitamins like D and A, accumulate in the body and can harm you in excess. Dose and form matter as much as the choice of product.

Not medical advice

This is informational coverage, not medical advice. Supplements interact with medications and conditions, and "over the counter" does not mean "risk-free for you." If you take any prescription medication, are pregnant or trying to conceive, or have an existing health condition, talk to a doctor or pharmacist before starting anything. Prescription products are prescription-only for a reason and are not part of this list.

A practical buyer's checklist: identify the actual gap first, prefer food, choose products carrying NSF Certified for Sport or Informed Sport, ignore "natural" and proprietary-blend marketing, and never assume more is better.

Sources

The Catalyst Feed

Content TeamIndependent, hands-on coverage of health, fitness & the tech that tracks it. Reviews you can trust — no hype.

#supplements #creatine #vitamin-d #omega-3 #protein-powder #evidence-based #third-party-testing