- This study will get tons of press, with misleading headlines all over the place. Like this one we already saw: "Low carb diets shorten your life unless you are mostly vegetarian."

- Nobody actually knows which diets are best for longevity, as it's an impossible question to answer without a trillion dollar randomized trial. Supercentenarians eat a wide variety of diets, calorie/methionine restriction helps in some animal studies, etc. A collection of observational evidence (like in the Lancet paper) is helpful for noticing overall patterns, but can't say much about actual diets. One high-carb diet could be high in candy and cake, and another high in fruit and tubers. The Lancet study wasn't able to address diet quality.

- The Lancet paper is not one study, but two: a long-term observational study in the US, plus a meta-analysis of international cohorts.

- The "easy" conclusions (low carb bad, high carb bad, high meat bad) are not so easy. The broadest way to cut dietary data is by macronutrient, so the result interpretations aren't always useful. Those with high and low carb intake are more likely to be on a diet, for example, and dieters can be more likely to have health conditions they're trying to address.

- When you meta-analyze heterogeneous data sets, problems arise. There's no good way around this, since the different data sets are from different research groups, and individual-level meta-analysis is thus rarely possible.

[1] https://www.facebook.com/Examinecom/posts/2146299502057199?_...