OTOH I'm wondering how you didn't break the LED during the reverse cycles of the AC wave, if you didn't use either a high-voltage diode in sequence or some random diode or second LED in reverse direction in parallel to the first one to bypass the other half of the wave.
My line capacitor suggestion was really just to get one started thinking about an actual circuit design for longer term. A 1/4 W resistor in serious with line voltage works fine on the test bench, but for longer term non-attended usage I would think about the various ways things could go wrong. And if you're going to spec a special resistor, then you might as well spec a special capacitor instead.
If I had to guess about why the LED wasn't destroyed, it's because the reverse breakdown voltage is low enough that significant heat does not build up in the LED. A reverse avalanche isn't necessarily destructive.