The same study says that low BMI is also associated with lower lifespan, and the statistically optimum BMI for white people is 23 to 25 (there are racial differences in all the numbers, which I think also challenges the idea of assigning moral qualities to them).

However, I don't know if I would give too much thought to the effect for low BMI - it could just be that people with extremely low BMI tend to have some disease process which causes that; only a small minority of them are rail-thin athletes. For high and extremely high BMI, there are known causal mechanisms for disease, so it's a different story.