Your citation measures 15-year mortality in adults 20-49. The P values for BMI's relationship with all-cause mortality and cardiac mortality were 0.071 and 0.030 respectively. It compares BMI against waist circumference and body fat percentage and suggests that the latter measures are better. I think it's misleading to say that "weight is not a primary predictor of health" based on this evidence.
First, the paper is talking about BMI rather than weight.
Second, what most people mean by "weight" in ordinary conversation is closer to body fat percentage than it is to BMI: Arnold Schwarzenegger was famously obese by BMI, but anybody who called him overweight during a conversation at the pub would likely be told he doesn't count.
Thirdly, the paper was close to statistical significance, even looking at young people and even with a cohort of a bit under 5000 people, so it doesn't rule out a correlation with BMI either (although yes, it does suggest BMI is a proxy for body fat, but this isn't a controversial statement).
Fourthly, GLP-1 agonists do reduce body fat[0], and body fat is the measure suggested by the paper you cited as being better than BMI.
First, the paper is talking about BMI rather than weight.
Second, what most people mean by "weight" in ordinary conversation is closer to body fat percentage than it is to BMI: Arnold Schwarzenegger was famously obese by BMI, but anybody who called him overweight during a conversation at the pub would likely be told he doesn't count.
Thirdly, the paper was close to statistical significance, even looking at young people and even with a cohort of a bit under 5000 people, so it doesn't rule out a correlation with BMI either (although yes, it does suggest BMI is a proxy for body fat, but this isn't a controversial statement).
Fourthly, GLP-1 agonists do reduce body fat[0], and body fat is the measure suggested by the paper you cited as being better than BMI.
[0] https://www.sciencedirect.com/science/article/abs/pii/S00142...