RF Enclosures

While I suppose it's possible this could detune the antenna I haven't heard any reports of such.  The manufacture makes no note of avoiding metal around it (in fact metal exists inside the unit) --- The core 50 Ohm parts are inside and away from the case by a bit. 

Commercial ISP's (those running thousands of units) swear by then, even if they are de-tuning they are still seeing improved link quality (so the detuning could not be significant)

Inside a NanoStation M2 it isn't a free space antenna. It is dual circuit boards (one providing the reference ground to which the radiating elements are tied)

After that we get to "failure" by detuning.  This would be a heat generated issue (reflected power.) The top Ubiquiti device puts out around 600mw.  That is minimal heat to be generated.  This is significantly less then a normal mobile puts out, we're not trying to throw back 50w of power into a small brick were talking at worst 300mw per chain (300mw*2=600mw). Factor in that anything more than 3db loss would probably shownup by causing issues and your down to a max of a VERY manageable 150mw heat disappation.  Considering sunlight can induce far more heat this seems to me to be insignificantly low to cause a failure.

These are are also the same types of chips used in home routers, those devices get abused left and right (including no antenna at all) and still keep kicking because the power is so low.