Hacker News Reader

Comment by greentxt | Hacker News Reader

"We found little or no effect of reducing saturated fat on all‐cause mortality (RR 0.96; 95% CI 0.90 to 1.03; 11 trials, 55,858 participants) or cardiovascular mortality (RR 0.95; 95% CI 0.80 to 1.12, 10 trials, 53,421 participants), both with GRADE moderate‐quality evidence.

There was little or no effect of reducing saturated fats on non‐fatal myocardial infarction (RR 0.97, 95% CI 0.87 to 1.07) or CHD mortality (RR 0.97, 95% CI 0.82 to 1.16, both low‐quality evidence), but effects on total (fatal or non‐fatal) myocardial infarction, stroke and CHD events (fatal or non‐fatal) were all unclear as the evidence was of very low quality. There was little or no effect on cancer mortality, cancer diagnoses, diabetes diagnosis, HDL cholesterol, serum triglycerides or blood pressure, and small reductions in weight, serum total cholesterol, LDL cholesterol and BMI."

This is your citation for saturated fat is evil?

View on HN

I don’t believe saturated fat is evil. I believe that consuming more than 10%E from SFA increases risk of CVD, and replacing SFA with PUFA reduces risk of CVD.

And yes, that is my citation in support of that claim. Quote: “There was a 17% reduction in cardiovascular events in people who had reduced SFA compared with those on higher SFA”

and

“When we subgrouped according to replacement for SFA, the PUFA replacement group suggested a 21% reduction in cardiovascular events”

I explain why the null findings on insensitive endpoints aren’t the gotcha some people think they are here: https://news.ycombinator.com/item?id=41960046

> There was a 17% reduction in cardiovascular events in people who had reduced SFA compared with those on higher SFA

they have table with breakdown of events by type, and in both rows with significant reduction it says that quality of evidence is "very low".

Don’t know where you’re seeing that. Combined cardiovascular events has an RR of 0.83 and the GRADE quality is moderate, not very low.

Yes, and I am talking about subtypes of events in the "summary of findings" table. Those subtypes with "moderate" quality has very insignificant diff in # of events. Those with significant diff in # of events have "very low" quality of evidence.

How would that affect the inference being made here? That is, the claim being made is with regard to combined cardiovascular events. This effect is significant and the GRADE quality is moderate.

One of the main purposes of meta-analyses is to summate multiple studies that may have insufficient statistical power in isolation to find a significant effect, but when summated they do.

I _think_ you're trying to suggest that the insignificant individual event types aren't contributing to the significant findings on the combined event endpoint, so we should ignore the GRADE quality of those (correct me if I'm wrong). But that's not how meta analysis functions - those events are also contributing to the significant finding on that endpoint (which is also why the GRADE quality of the combined endpoint is moderate). They just aren't significant _in isolation_.

> How would that affect the inference being made here?

to me, it looks like several rows clearly don't add up in key findings table(they derive "moderate quality" conclusions from "very low quality" observations), which makes whole publication as data astrology quality.

Right, but that’s because you’re misreading the table. They are deriving moderate quality conclusions from moderate quality data, as I’ve explained.

Looks like we are in disagreement about who is misreading the table.

Well, good luck sustaining the claim that Lee Hooper doesn’t know how to perform a meta analysis and that error got past Cochrane’s notoriously lax peer review process, and you definitely didn’t just make a mistake reading the table.

I salute your self confidence, I must say I’m somewhat envious of it.

I have low opinion about modern academia and quality of "peer review processes".

If you misread data in the way you appear to, that doesn’t surprise me. I can’t imagine you think many studies’ conclusions appear to follow from their results.

It’s just that you’re blaming the studies, rather than your comprehension.

Again, we are in disagreement who misread the data.

Honestly, your steadfast insistence that you couldn’t possibly be wrong on this is fascinating to me. I’m happy to admit I made a mistake if I’m misreading the data, so walk me through what you think I’m getting wrong.

Please be specific: name the event subtypes you’re looking at, give specific references to the table numbers etc. If I’m wrong about this, I want to know.

Meta-analyses are a cool way to perform a garbage in, garbage out exercise.

What’s the evidence that the data going in is garbage? We already ascertained most of it is GRADE quality moderate.

I don’t research the topic for a living, so I don’t know — I’ve read none of the studies.

If the conclusion that comes out of the other end of the pipeline is that we should avoid eating the things we’ve eaten for millions of years, it’s fair to say what went in might have been based on false premises.

Why? Evolution doesn’t select for longevity, it selects for reproductive fitness. We have no reason to believe ancestral foods are more longevity promoting than novel ones, and antagonistic pleiotropy would suggest that all else held equal, ancestral foods are less healthy.

Heck, water treatment plants are novel compared to the pathogen-laced water we’ve been ancestrally consuming, but I doubt you eschew tap water for puddle water.

Reply on news.ycombinator.com