Pseudo Science

This dataset is public. We can open it up and see that age-adjustment kills this 'finding'. This is totally clinically meaningless: It's a <0.5% increase in fracture risk, it's less than the effect of one year of post-menopausal bone loss, it's below DXA measurement error!

Do Statins make you Fat? Got your attention. Good. Now hear me out. 👂👇 The new headlines around this new Lancet study screaming that “millions more” should be on statins and that “side effects have been overblown” are absurd distortions of reality. Take weight change (Figure 1). There was a clear trend showing statin use was “causally” linked with weight gain versus placebo (missed statistical significance by a hair; yellow line). AND there was also a dose–response trend (Figure 2): high-dose statins trended toward more weight gain than low-dose statins (also missed statistical significance by a hair). 🚨Here’s the problem: Just because something doesn’t cross an arbitrary p-value threshold does not mean there’s no biological effect. If I’m 90% sure jumping off a cliff will kill me—but that’s “not statistically significant”—does that mean it’s safe? Of course not. Absence of statistical significance ≠ absence of harm. (And maybe it’s not my place to speculate on whether conflicts of interest—including with Merck, AstraZeneca, Eli Lilly, Novartis, Amgen, Pfizer—influenced which tests they ran and what they chose to emphasize... but FWIW to you 🤷🏼) Many more thoughts in the full letter 👇

Chemistry is everywhere. You can reject it, complain about it, or fear it…but it’s literally the reason you are alive.