Hacker News Reader

Comment by tohnjitor | Hacker News Reader

FDA approval is mostly meaningless. Just this year they approved an Alzheimer's treatment which ended clinical trials early because it didn't work.

https://blogs.sciencemag.org/pipeline/archives/2021/06/08/th...

View on HN

I think you're over-simplifying the Alzheimer's treatment. The medicine was shown to be effective in reducing plaques, and has been approved for any medical use where reducing plaques is the goal. However, reducing plaques does not seem to actually help patients with Alzheimer's. So the medicine "works" in the sense that it does what it's supposed to (reduce plaques) but does not help Alzheimer's patient outcomes and should not be used for that purpose. The FDA made the bone-headed decision to let the medicine get its seal of approval anyway, knowing full well that every doctor in the USA would prescribe it anyway.

> However, reducing plaques does not seem to actually help patients with Alzheimer's

Just like reducing cholesterol does not lower heart attacks.

That article uses Gary Taubes as a reference.

I like Gary, but on the issue of cholesterol is more complex than he portrays.

It is absolutely the case that LDL-C is correlated with heart disease at the population level. It is absolutely the case that LDL-P is correlated with heart disease at the individual level. Beyond the correlation, there is considerable mechanism of action evidence that small particle lipoproteins are the driver of atherosclerosis.

The primary issue with statins / LDL is that they are prescribed based on LDL-C, which is not appropriate at the individual level and are thus considerably over proscribed. LDL-C only matters at the population level, not individual. When lowering LDL-C for the population, you will only see small benefits since most people get little to no benefit from the drug.

However, people with high LDL-P will absolutely see a benefit from statins. You need only look at the FH groups to see this.

To look at that Alzheimer's approval and then conclude that FDA approval is mostly meaningless is a pretty strong statement based off of that alone. Even if you add in high profile drugs that were approved and later were shown to cause huge problems, it still doesn't make FDA approval meaningless. It's pretty hard to get a drug to be FDA approved and means quite a bit.

"mostly meaningless" is an overstatement. The FDA oversees a lot of drug applications and gets it right a vast majority of the time. No human endeavor is perfect; if there's evidence that 20% of approvals are wrong then I'll agree that there's a problem.

>No human endeavor is perfect

This should be posted to the front page of the site. I get so tired of seeing comments pointing out some flaw in an institution and concluding that the institution is therefore useless. It really is lazy thinking.

It is also lazy thinking to say that since it is approved by an institution it must be good.

I think the point is more about putting "authority" into perspective.

One third of all FDA approved drugs from 2001 to 2010 were pulled from the market post-approval.

So it’s not 20%. It’s a third.

The approval for that drug was a huge error but you’re incorrectly conflating that limited, conditional approval with the far more rigorous normal process. The FDA basically said that the manufacturer has 10 years to show a clinical benefit because the drug was shown to accomplish its purpose (reducing amyloid plaques) with an acceptable level of side effects — and that’s not completely off the timeframe needed to measure a slower disease like Alzheimer’s. In contrast, the Pfizer vaccine has been shown to be highly effective at producing positive clinical outcomes and has been administered to massive populations with very infrequent side-effects which are far below the risks of the disease.

I think it was a big mistake not to keep that as some kind of trial but my primary reason for that is because it encourages faulty conflations like yours. The message we should take from this is that the problem was NOT following the standard process rather than drawing any conclusions about the process which wasn’t followed.

No, it’s not meaningless. That’s hyperbole and a misunderstanding of what’s going on.

That story was all over the headlines not because it was normal. It was in headlines, and in the front of your mind, because it was the exception that proves the rule. That’s not typical.

Regardless, the FDA approval was mostly about assessing safety of the drug in controlled trials and follow-ups. Likewise, the vaccine studies were rigorous examinations of safety and an in-depth search for warning signs that might indicate problems in small populations. The correct interpretation of this approval is to see that it has been thoroughly examined for safety in the field.

Finally, we already have mountains of evidence that the vaccine is effective. That much is not in question after administering an incredible number of doses globally and observing the outcomes. Proving efficacy was the least of the FDA’s concerns at this point.

> That story was all over the headlines not because it was normal.

It was in the headlines because it is normal, and some people were trying to improve the system but they failed. There are tons of existing drugs that have been approved based on surrogate endpoints, e.g. iirc the entire class of statin drugs.

This suggests FDA approval doesn’t say much about whether something works—presumably it still says something about safety. At any rate, whether or not FDA approval means much, to a certain subset of people it ostensibly did, since the objection that there was only an EUA was fairly common.

If the treatment is ineffective and carries any side effects, then it is a safety issue. The whole "do no harm" principle means that you are subjecting a patient to unnecessary risks and side effects with no possible benefit.

This is really u fortunate and true. Approving a drug that doesn’t work and rushing a vaccine through the process may have the effect of less faith in governmental organizations.

Reply on news.ycombinator.com