Search results for CBD promise relief from nearly everything, and the fine print rarely keeps up. The reality is more mixed: one use has strong trial support, a handful of others have early signals, and the list of reported side effects is real and fairly consistent. Sorting CBD oil into those three piles is the quickest way to read any health claim.

What the Evidence Supports

One approved use

In June 2018 the FDA approved Epidiolex, a purified CBD oral solution, for seizures linked to Lennox-Gastaut syndrome and Dravet syndrome, two rare and severe childhood epilepsies. Tuberous sclerosis complex was added in 2020. The approval rested on randomized, placebo-controlled trials.

In a 2017 trial in the New England Journal of Medicine, 120 children and young adults with Dravet syndrome took CBD at 20 mg per kilogram of body weight per day. Convulsive seizures fell by a median of 38.9 percent, compared with 13.3 percent on placebo. A 2018 trial in The Lancet followed 171 patients with Lennox-Gastaut syndrome. Drop seizures fell by a median of 43.9 percent with CBD and 21.8 percent with placebo.

Those numbers come from a pharmaceutical-grade product, a high dose, and medical supervision. They show that CBD can work as a medicine for specific conditions. They do not show that a bottle bought online works for anything else.

Why placebo groups matter

The placebo numbers deserve as much attention as the CBD numbers. In the Dravet trial, seizures fell by 13.3 percent on placebo, which means the net difference attributed to CBD was about 25 percentage points. In the Lennox-Gastaut trial the gap was roughly 22 points. Seizure counts are objective, so those gaps are real effects. For conditions judged by feel, such as pain, anxiety, and sleep, placebo responses run larger and noisier, and a trial without a placebo arm can make almost anything look helpful.

Duration and dose

Both epilepsy trials lasted about 14 weeks, and the dose was tied to body weight and adjusted under medical supervision. Retail CBD oil comes in bottles labeled by total milligrams, with serving suggestions that are usually a small fraction of the trial amounts. That difference does not make retail oil useless, but it does mean trial results cannot be borrowed for a bottle off the shelf.

Everything else

Use	Best available evidence	Strength
Certain childhood epilepsies	Large randomized trials of Epidiolex	Strong
Anxiety in stressful situations	Small placebo-controlled trials using single doses of 300 to 600 mg	Limited
Sleep	Case series and small studies without strong controls	Low
Chronic pain	Mixed trials, including a 2022 trial of 20 to 30 mg daily that matched placebo	Low
Acne and other skin conditions	Lab studies and small human reports	Early
Most other claims	Animal or cell experiments	Insufficient

Most health claims sit in the bottom rows. Cell and animal studies are good at generating ideas, though many promising compounds fail once they are tested in people. A claim that cites only a lab result is a hypothesis.

Benefits people report versus benefits measured

Surveys describe relaxation, better sleep, and less pain. Those are real experiences, though self-reports are shaped by expectation, product variation, and the natural ups and downs of daily life. Trials try to strip those influences away. When a trial and a survey disagree, the trial carries more weight, and the survey still points researchers toward questions worth testing.

Plant origin and safety

Plant-based does not mean harmless. Digoxin, a heart medicine, comes from foxglove, and it has a narrow safety margin. CBD sits somewhere much safer than that, though it is still a pharmacologically active compound that interacts with enzymes, receptors, and other drugs. Treating it like a food with no downside overlooks the same trial data that supports its benefits.

What Side Effects Look Like

The common ones

A 2020 meta-analysis in Neuropsychopharmacology pooled 12 randomized trials with 803 participants. Abnormal liver tests, sleepiness, sedation, and pneumonia showed up mainly in childhood epilepsy studies, where CBD was taken alongside other seizure drugs such as clobazam and valproate. With those studies set aside, the only outcome tied to CBD was diarrhea, with an odds ratio of about 5.

Later reviews lengthen the list a little. A 2022 update of randomized trials of oral CBD reported these effects across studies:

• Diarrhea
• Drowsiness and sedation
• Decreased appetite
• Dizziness and headache
• Nausea, vomiting, and abdominal discomfort
• Rash

Most were mild or moderate. Serious adverse effects were uncommon, and many of them occurred in people taking several other medicines at once.

The liver

Liver enzymes get the most attention. The Epidiolex prescribing information reports that alanine aminotransferase (ALT) rose above three times the upper limit of normal in about 13 percent of patients taking 20 mg per kilogram daily, compared with about 1 percent on placebo. The rate was higher in patients also taking valproate. Before starting the drug, doctors check liver blood tests and repeat them after one, three, and six months.

The FDA has also named liver injury as a concern for nonprescription CBD. Retail doses are usually far lower than the trial doses, though products vary and no one tracks the long-term effect of daily use at those amounts.

Interactions

CBD is processed by liver enzymes, including CYP3A4 and CYP2C19, and it can slow the breakdown of other drugs. In epilepsy care it raises the level of an active byproduct of clobazam, so doctors sometimes lower that dose. Case reports describe higher INR readings, a measure of blood clotting, in people who combined CBD with warfarin. Alcohol, sleep aids, and other sedating medicines can add to the drowsiness CBD causes on its own.

Weighing It Up

Three questions before buying

A short checklist keeps the decision grounded:

1. What problem is being treated, and does a better-supported option already exist? For sleep, anxiety, and pain, structured therapy and habit changes have far more evidence behind them.
2. What else is being taken? Prescription medicines, supplements, and alcohol all count.
3. Who is taking it? Age, pregnancy, and liver health change the risk.

The first question carries the most weight. Poor sleep, for example, often has a cause that CBD does not touch, such as late caffeine, a noisy room, or an irregular schedule. Fixing the cause costs less and works better than adding a supplement on top. The same logic applies to anxiety and pain, where structured therapy and exercise have strong track records.

Who needs extra care

• People who take prescription medication, especially blood thinners and seizure drugs
• Anyone with liver disease or heavy alcohol use
• People who are pregnant or nursing
• Older adults, who are more sensitive to sedating effects
• Anyone who drives or operates machinery for work

Product quality

Side effects depend on what is actually in the bottle. A 2017 JAMA analysis of 84 online CBD products found THC in about one in five, and potency often missed the label. A product that causes unexpected drowsiness may contain more CBD than the label states, or a trace of THC, or sweeteners and flavorings that upset the stomach. A batch-specific lab report from an independent lab reduces those surprises.

Reading health claims with a sharper eye

Health claims about CBD vary widely in quality. A few quick checks sort the careful ones from the careless ones:

• Does the claim come from a human trial, or from cells and animals
• Was there a placebo group, and were participants randomly assigned
• How many people took part, and for how long
• Was the dose reported, and does it match what a bottle delivers
• Does the claim name a specific condition and outcome, or promise broad wellness

Words such as “clinically proven” and “miracle” are warning signs. Real findings usually sound modest, with numbers attached and limits stated.

Price is a poor stand-in for quality. Expensive bottles can come with weak testing, and modest ones can come with thorough reports. The lab report, not the price tag, is the evidence.

A sensible way to start

Choose a product with a lab report, begin with a small serving, and change only one thing at a time. A simple note with the date, serving, and any effects makes patterns visible. Stop and contact a clinician for yellowing of the skin or eyes, dark urine, persistent nausea, severe fatigue, or a rash with swelling.

Doctors and pharmacists are used to these questions. Bringing the bottle and the lab report to the conversation saves time and gets better advice.

Where the research is heading

The 2020 meta-analysis on adverse effects noted that more safety data is needed from trials outside childhood epilepsy and from studies of over-the-counter CBD products. That gap matters, because most trial data comes from a pharmaceutical product in a narrow patient group. In 2023 the FDA told Congress that existing food and supplement rules do not fit CBD and that a new regulatory pathway is needed. Until better data and clearer rules arrive, caution and good records are the practical tools available.

Final Thoughts

CBD oil has strong evidence for a few rare childhood epilepsies and thin evidence for most other uses. Side effects are usually mild, with diarrhea and drowsiness leading the list, while liver changes and drug interactions deserve real attention. Product quality adds another layer of uncertainty. A short chat with a pharmacist and a notebook of effects make the benefits and risks much easier to judge.