DependsThis is essentially the question of whether a $k$-unirational variety is necessarily $k$-rational. The short answer is No. The following longer answer mostly summarizes some of the exposition at http://en.wikipedia.org/wiki/Rational_variety; for more information see that page and the references it gives.
The existence of a non-rational subfield $F$ of $K$ depends on $k$ and $n$. If If $k$ is algebraically closed and of characteristic zero, then then the answer is No for $n=2$ by a theorem of Castelnuovo Castelnuovo (and for $n=1$ by a theorem of Lüroth), but but Yes for $n=3$, and presumablythus for all $n \geq 3$ (you did not require $F/K$ to be a finite extension). In positive In characteristic $p>0$ things can get much stranger: Zariski gave examples for $n=2$ where the extension $K/F$ is inseparable; and more recently Shioda constructed, for each $n \geq 2$ there are inseparable extensions and every power $K/F$$q$ of $p$, an example where $F$$K/F$ is the function field of a variety of general type. Shioda gave explicit examples whereinseparable and $F$ is the function field of the the Fermat hypersurface of degreedimension $q+1$$n$ and degree $q$$q+1$ (which is any power of the characteristic;general type once $q \geq n+3$), see Propositions 1 1 and 3 ofin
Shioda, T.: An Example of Unirational Surfaces in Characteristic $p$, Math. Ann. 211 (1974), 233-236.
The exposition and references at http://en.wikipedia.org/wiki/Rational_variety may also be of interest.