Understanding Evidence
Why do animal studies and human studies give different answers?
TL;DR — What the Evidence Shows
Most positive CBD evidence comes from animal and lab studies. Human trials for arthritis pain are mostly negative. This gap is the central challenge when evaluating CBD.
What do animal studies show?
Several animal studies have found that CBD reduces inflammation and pain-related behaviors in rats with arthritis. These results look promising at first glance.
Hammell 2016: Transdermal CBD (applied through the skin) reduced joint swelling and pain-related behaviors in a rat model of arthritis. Researchers found lower levels of inflammatory markers in the treated joints.
Philpott 2017: CBD prevented pain and nerve damage in a rat model of osteoarthritis. It also reduced joint inflammation (swelling of the tissue lining the joint).
Lab studies (cell culture research) also show that CBD can reduce inflammatory signals in isolated cells. These include suppression of TNF-alpha and IL-6 (proteins that drive inflammation) and inhibition of NF-kB (a pathway that controls inflammation).
Multiple lab studies show CBD suppresses inflammatory markers in cell cultures. These findings tell us about biology, but cells in a dish do not behave the same as cells inside a living person.
Nowell 2024; Campos 2023; Costa 2004.
What do human trials show?
When researchers tested CBD for arthritis pain in controlled human trials, the results were mostly negative. Three completed randomized controlled trials specifically tested CBD in people with osteoarthritis. None found a significant benefit.
Vela 2022 (136 people): Oral CBD for hand osteoarthritis and psoriatic arthritis. No significant difference from placebo on the primary pain outcome. The difference was 0.23 mm on a 100 mm pain scale.
Bialas 2023 (86 people): Oral CBD added to paracetamol (acetaminophen) for knee osteoarthritis. No significant difference from placebo on the WOMAC questionnaire.
CANOA / Mojoli 2025: CBD-rich cannabis oil for knee osteoarthritis. No significant difference from placebo on the WOMAC pain questionnaire.
Looking beyond arthritis, a major review of 16 randomized controlled trials of CBD for any type of pain found that 15 out of 16 were negative (Moore et al. 2024). A U.S. government systematic review (AHRQ 2025) reached a similar conclusion: CBD alone was not associated with decreased pain intensity.
The score so far
- Animal studies: positive results for inflammation and pain behaviors
- Human trials for arthritis pain: 3 out of 3 negative
- Human trials for any pain: 15 out of 16 negative
Why the gap?
There are several reasons why a treatment that works in animals may not work in humans. For CBD, three factors are especially important.
bioavailability is very low
When you swallow CBD, only about 6 out of 100 milligrams actually reach your bloodstream. (Huestis 2019) The rest is broken down by your liver before it can have an effect. This means the dose that reaches your joints may be far lower than the dose that worked in cell cultures or was injected directly into rat joints.
Doses do not translate directly
In animal studies, researchers can deliver precise doses directly to the target tissue. In human trials, CBD must survive digestion, pass through the liver, travel through the bloodstream, and reach the joint. The dose that reduced inflammation in a rat knee does not predict the oral dose a person would need for the same effect.
The outcomes measured are different
Animal studies measure inflammation markers and pain-related behaviors (like how much a rat limps). Human trials measure patient-reported pain, which is what people actually care about. It is possible for a treatment to reduce inflammation markers without reducing the pain a person feels.
These problems are not unique to CBD. Across all of medicine, only about 1 out of 20 treatments that show promise in animals eventually receive approval for use in humans (Ineichen et al., PLOS Biology, 2024). The failure to translate is the rule, not the exception.
What does this mean for claims you will see?
News articles and product websites often describe CBD research without distinguishing between animal and human findings. A headline like "CBD reduces arthritis inflammation, study finds" may refer to a rat study. This is not wrong, but it is incomplete.
When evaluating a CBD claim, the most important question to ask is: Was this tested in people?
- If the evidence is from human trials, it directly answers whether CBD works for people
- If the evidence is from animal studies, it tells us about biology but does not prove human benefit
- If the evidence is from lab studies (cell cultures), it is the furthest from real-world treatment effects
How do we label these on this site?
Throughout this site, every claim carries a source type tag. Here is what each one means.
| Tag | What it means | Can it get a GRADE framework rating? |
|---|---|---|
| Human Trial | Tested in a controlled study in people | Yes |
| Human Survey | Self-reported by users, no placebo control | Generally no (cannot distinguish from placebo effect) |
| Animal Study | Tested in animals (usually rats or mice) | No — receives "Not Rateable" designation |
| Lab Study | Tested in cells in a dish | No — receives "Not Rateable" designation |
When you see an animal study or lab study tagged on this site, it will never carry a clinical certainty badge. This is intentional. Giving preclinical findings a certainty rating would suggest they apply to human treatment, which has not been shown.
Key sources cited on this page
Page last reviewed: March 2026 · Authored by Claude (Anthropic AI) · Research methodology