BMI is not a valid statistic. Anyone who mentions BMI invalidates any other post they might ever make or have made.
It's not as bad as many people think. The scaling exponent is wrong, but in the range of values that is relevant it works ok. Longitudinal studies indicate that people with BMIs over 25 or over 30 or over 35 are at progressively greater risk for heart disease and diabetes and other medical problems.
A lot of people stupidly interpret "overweight" as "fat" and so say things like "but I am 6 foot of pure muscle and not fat although I have a BMI of 30". What they don't understand is that a 6 foot 210lb guy who is all muscle has an elevated risk of health problems, like heart disease. All that muscle and weight isn't healthy.