I would be pretty surprised if there was a non-calculus proof of this that wasn't just "calculus in disguise", because you need to look at higher order behaviour of the trig functions (hence the differentiations).
Inequalities like sin(x) < x lose too much information.