Why I Think Tylenol is Both Dangerous & Useless

This post contains a video, which you can also view here. To support more videos like this, head to!

I learned something important last week: I learned that you, the gentle viewers of this channel, really, really love your painkillers. Maybe too much? Who am I to say!

So, yes, in my video on “rainbow fentanyl” last week, I dropped a little aside about how Tylenol, aka acetaminophen, aka paracetamol, is a trash drug that is no better than placebo and has the bonus effect of killing about 500 Americans per year. This is a little nugget of information I do like to share whenever relevant because it’s something that not a lot of people know, and I know I was appreciative when I learned about it. But you guys! Many of you were NOT appreciative, and in fact you were very sure that I was absolutely wrong:

“Mjl” writes “it’s a really useful drug clinically, i’m not sure where the “trash drug” notion comes from.

yes, it’s toxic in high doses, but a lot of drugs are. it’s a lot better and safer than ibuprofen and other mild pain relievers”

And “sunyavadin” writes “The problem for its effectiveness is when it’s used for stuff it’s not indicated for, if you need a broad over the counter analgesic, Ibuprofen all the way. If you need something specifically to target for example, acute migraine symptoms, you want Paracetamol. Its specificity is why it won’t work for things like inflammatory pain, musculoskeletal pain, etc. I think it’s mostly a problem of confusing people over what it’s for with marketing. It is however useful for enhancing the effects of other analgesics taken in combination with it, due to its interactions with them, so there’s some use for it if used correctly.”

Farimira says “It feels effective to me even on supposed inflammatory pain. I’ve been prescribed it for a fever by a doctor once because it can help with inflammation…. I’ve heard that it doesn’t work from a medical scientist but I think perhaps the mechanism is not understood and needs more research. I reckon in this case placebo is used to handwave away something we don’t understand.”

That’s just a small sample, alongside some very polite requests for me to do a video about it. So! This is that video. I regret to inform you that all the commenters I just read out are wrong.

First, let me just point out that something “working” for you does not mean it can’t be a placebo. That’s part of the definition of a placebo! Let’s go over it really quickly with an imaginary situation: let’s say I have a pill that I think can fix a headache. So I take 200 people with headaches and I give 100 of them my pill. I give the other 100 people a sugar pill. Then I wait two hours and ask them if their headaches cleared up.

59 of the 100 people who took my pill say it absolutely worked for them! Their headaches are gone! That’s not bad, is it? My pill helps more than half of the people who take it. (That’s you, by the way, you were in that group and you are now a lifelong customer for my pills, thank you for your business.)

But wait, what about the group that took the sugar pill? Well! 49 of 100 people in THAT group say the sugar pill helped their headache go away. They love it, and are wondering where they can go buy more for future headaches. Huh. That means my pill DID do better, but only by 10 people out of 100. So, instead of my pill being 59% effective, it’s actually only 10% effective.

Make sense? My pill “worked,” but only a little better than placebo. The people who took the placebo felt just as much that their pill “worked.”

Now for a fun reveal: I lied. I didn’t make that example up. Those are the actual statistics found in a Cochrane review that tried to find out if Tylenol/acetaminophen/paracetamol actually helps with tension headaches. I’ve talked about Cochrane before but essentially they are a British nonprofit that provides gold standard systematic reviews and meta-analyses that serve to sum up the current science on a particular topic. And yeah, in this case they found that only 10 out of 100 people saw any benefit that could be reasonably attributed to Tylenol, across 23 studies that actually included more than 8,000 subjects. Ibuprofen, for the record, only helped 7 out of 100 people. So. Also not great!

That’s just tension headaches, of course. Tylenol might do a little better for people with migraines, at least according to studies funded by McNeil Consumer Healthcare, the Johnson & Johnson Subsidiary that makes Tylenol. I’m not being cute, here, I really had trouble finding good studies specific to migraines that were done by anyone other than McNeil Consumer Healthcare. But yes, according to a McNeil Consumer Healthcare paper, 52% of patients who took Tylenol got migraine relief compared to 32% of those who got placebo.

I DID eventually find a systematic review that came from the Mayo Clinic, which found that Tylenol offered a small potential benefit though they continue to recommend triptans (migraine-specific pain relievers) and NSAIDs (Non-Steroidal Anti-Inflammatory Drugs like aspirin, ibuprofin, and naproxen – technically, Tylenol is an NSAID but it’s so weak that it isn’t categorized as such) as being the first line of defense for a migraine for most people, since they have a greater effect with fewer immediate adverse side effects (though they also specifically point out that Tylenol is, long-term, toxic to your liver, which I’ll get to in a bit).

Another systematic review looked at Tylenol in combination with aspirin and caffeine, a commonly used cocktail for headaches. They found that that combo DID perform better than placebo across several studies, though as you can see it varied wildly how MUCH better it performed: four times better in a few early studies, or just a tiny bit better in later studies that found 35% of people getting better after two hours with Tylenol compared to 26% of placebo.

So it might work for migraines, but probably not as well as the drugs that specifically are made to help migraines. That’s why one British Medical Journal editorial described Tylenol as having “modest efficacy” in treating migraines and headaches.

And that, I’m sorry to say, is where the good news for Tylenol ends. Yes, that was the good news. The rest is pretty bleak. To quote from that same editorial:

“Paracetamol at doses between 500 and 1000?mg is in the least effective quartile of drugs for treating acute postoperative pain.

Paracetamol at doses up to 4000?mg daily is ineffective in back pain.

Paracetamol at doses up to 4000?mg daily is practically ineffective in arthritis. Though marginally better than placebo, paracetamol has little chance of achieving clinically meaningful benefit in osteoarthritis.

No review evidence that paracetamol works for dysmenorrhoea (period cramps), neck pain, rheumatoid arthritis or cancer pain.”

This has been shown in review after review: Tylenol/acetaminophen/paracetamol is just not very good at helping various types of pain. Toothaches, back aches, surgical pain, colds: you may as well just have a sugar pill.

In fact, you’d be better off taking the sugar pill because it’s safer. As I stated in the previous video, Tylenol is responsible for 500 American deaths per year, but there’s way more: Tylenol overdoses are also responsible for 50,000 emergency room visits and 25,000 hospitalizations each year. The only reason why the number of deaths isn’t higher is because quick treatment can allow a person to live even with severe liver damage.

But you don’t have to OD on Tylenol for it to do damage:

“In clinical trials in chronic pain, patients taking paracetamol were four times more likely to have abnormal results on liver function tests than those taking placebo.”

“Acute liver failure leading to registration for transplantation was twice as common in non-overdose paracetamol-exposed patients than with NSAIDs in a large case-population study.”


“A systematic review of observational studies found that compared with people not taking paracetamol, paracetamol use, especially at higher doses, was associated with increased mortality, cardiovascular adverse events (fatal or non-fatal myocardial infarction, stroke or fatal coronary heart disease), gastrointestinal adverse events (gastroduodenal ulcers and complications such as upper gastrointestinal haemorrhage) and estimated glomerular filtration rate decrease of at least 30?mL/min/1.73?sq?m (meaning it fucked up the kidneys).”

“Therapeutic index” is a ratio that scientists use to compare a drug’s effective dose (how much you need to take to get the benefit) and its toxic dose (how much you need to take to severely injure or kill you). The further those numbers are from each other, the higher the TI and the safer the drug. For instance, the opioid remifentanil (not to be confused with its cousin, fentanyl) has a TI of 33,000, which is fantastic: you only need a tiny bit to feel better, but a whole helluva lot to OD. Meanwhile, cocaine has a TI of 15: the amount you take to get high is way closer to the amount you’d take to OD.

So, where on that range does Tylenol fall? Down here….below cocaine. Tylenol has a therapeutic index of 10. And that’s actually a pretty optimistic number, because I’ve seen some doctors pin it around 3 or 4, especially before they lowered the recommended dose in 2011.

All of this isn’t to say that other painkillers don’t have their own drawbacks: I can’t take ibuprofen and other NSAIDs because they tend to aggravate a painful chronic condition I have called interstitial cystitis, so when I have a headache or back ache I just, well, put up with it. Lots of people are allergic to some painkillers, and NSAIDs tend to aggravate the stomach. So please don’t start arguing in the comments about “well it’s better than Advil” or whatever. There is no universal “better.” There is only what is right for you, and your specific health concerns. I’m not a doctor and I’m not telling you to stop taking Tylenol or any other drug. If it’s working for you, if a doctor suggested it, great. I just want you to be informed about what the data actually says: if you’ve been taking this drug every day because you think it’s harmless, since you can just go into Costco and buy an entire bucket of it right off the shelf, maybe it’s time to talk to a doctor about whether it’s really helping you, because if it’s not, the downsides are pretty bad.

Rebecca Watson

Rebecca is a writer, speaker, YouTube personality, and unrepentant science nerd. In addition to founding and continuing to run Skepchick, she hosts Quiz-o-Tron, a monthly science-themed quiz show and podcast that pits comedians against nerds. There is an asteroid named in her honor.

Related Articles

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Back to top button