What supplements are actually worth taking?
The short list is boring: correct real deficiencies, consider creatine if you train, use protein or fiber when food falls short, and be cautious with anything marketed as anti-aging. The more magical the promise, the more I want a boring reason to use it.
Worth considering
Deficiency replacement, creatine with training, protein/fiber gaps, and selected symptom-specific trials.
Usually skip
Detox products, proprietary blends, mega-dose antioxidant stacks, and anti-aging bundles with no endpoint.
My rule
One supplement, one target, one timeline. If nothing changes, stop it.
The filter
A supplement has to pass one of three tests: it fixes a deficiency, it improves a measurable outcome you care about, or it has a plausible safety/benefit profile for a specific situation. If it only passes the influencer test, no.
The right question is not whether a supplement has a mechanism. Almost everything has a mechanism if you zoom in far enough. The question is whether a human using a real-world dose gets a benefit worth the cost and risk.
The short list I take seriously
This does not mean everyone needs these. It means these categories have a clearer reason to exist than the average longevity stack.
Deficiency replacementiron, B12, vitamin D, iodine, calcium, or other nutrients when diet, labs, bone risk, pregnancy context, or medications make the gap real.
Training supportcreatine monohydrate when resistance training or power output matters.
Diet supportprotein powder or fiber when the actual diet is short and food alone is not solving it.
Symptom-specific trialsmagnesium for constipation or selected sleep complaints, electrolytes for heavy sweating or low-carb transitions, omega-3 only when the indication makes sense.
Where supplements can hurt
The harms are not just liver injury headlines. The more common harm is quieter: people spend money, add complexity, ignore medications or proven prevention, and then chase borderline labs created by the stack.
Interactions matter. So do dose, form, quality control, kidney and liver function, pregnancy, anticoagulants, cancer therapy, transplant meds, and surgery timing.
How I would run a supplement trial
Pick one target and one timeline. If the target is ferritin, measure ferritin and symptoms. If the target is training, track lifts or sprint output. If the target is sleep, track latency, awakenings, morning grogginess, and whether caffeine/alcohol/light timing changed at the same time.
Do not start five things at once and then declare victory. That is how a supplement cabinet becomes a mythology project.
What I would not do
I would not take beta carotene or vitamin E for cardiovascular or cancer prevention. I would not buy detox products. I would not use a multivitamin to excuse a poor diet. And I would not take a blend that refuses to tell me the dose of each ingredient.
When to talk to your doctor
Talk first if you are pregnant, trying to conceive, on anticoagulants, taking cancer/transplant/immune medications, have kidney or liver disease, have unexplained symptoms, or are considering high-dose supplements. Bring the label and the dose, not just the brand name.
Clinical lens
How I’d decide
Use this section as a second pass after the main answer, not as homework before you know what the page is saying.
Who it’s for
People with documented deficiencies, dietary constraints, heavy training goals, pregnancy/perimenopause-specific needs, medication patterns that create predictable gaps, or a measurable symptom/outcome they are trying to change.
Who should skip it
Anyone on anticoagulants, cancer therapy, transplant medications, pregnancy-related restrictions, or complex prescriptions should review supplements clinically. Also skip proprietary blends, mega-dose stacks, and anything that hides the dose.
Measure before / after
Measure the deficiency or outcome you are targeting: ferritin/iron studies, B12, vitamin D when relevant, lipids, glucose, symptoms, training performance, sleep, GI tolerance, and medication interactions.
What I’d do first
Start with food, training, sleep, alcohol/smoking, and medications. Add one supplement only when there is a reason, dose it sanely, and stop it if the target does not move.
What would change my mind
I upgrade supplements when human outcome data show meaningful benefit and low harm. I downgrade them when benefits vanish outside small biased studies, contamination risk is meaningful, or the supplement mostly sells a story.
Frequently Asked Questions
Do I need a multivitamin?
Maybe, but it is not my default answer to longevity. It can be useful for specific dietary gaps, pregnancy planning, or restricted intake. It is not a substitute for protein, fiber, plants, sleep, training, or risk-factor management.
Which supplements are worth testing labs for?
Iron/ferritin, B12, and vitamin D are common examples when risk or symptoms fit. I would not order a large micronutrient panel unless the result has a clear action pathway.
Are supplement brands reliable?
Quality varies. I prefer transparent dosing and third-party testing when possible. I am more skeptical of proprietary blends, extreme claims, and products sold with a protocol that only the seller can interpret.
Can supplements replace medication?
Sometimes a supplement can support a plan, but it should not replace indicated treatment for high blood pressure, high ApoB, diabetes, thyroid disease, anemia, or another diagnosis without a clinician-guided reason.
References & citations
- 1.USPSTF recommendation: vitamin, mineral, and multivitamin supplementation to prevent cardiovascular disease and cancer
- 2.NIH Office of Dietary Supplements fact sheets
- 3.International Society of Sports Nutrition position stand: creatine supplementation
- 4.VITAL trial: vitamin D supplements and prevention of cancer/cardiovascular disease
Related Guides
Next step
Turn the guide into the right next decision.
If this page raised a real clinical question, start with the practice details. If you are still learning, get the weekly letter. If you are comparing tests, use the testing hub before buying another panel.