Yeah, the Odyssey, the Mekong, the Orinoco, and the Rio Grande were all retreating back to the Wormhole, away from the Jem'Hadar fleet, when the Jem'Hadar ship made its suicide run. They'd been defeated, but they weren't dead in the water or anything; they would have retreated to Deep Space 9 for repairs.
Except that the ship was not 'almost completely deprived of power'. We are talking about what is probably carefully contained and highly volatile energy (plasma) that runs through the ship on a continuous basis - and since the Feds have the ability of converting energy into matter, they can also convert energy from one form into another - this was corroborated by Torres when she state that they also 'recycle energy' - amounting to 0 waste. It is conceivable that they can in that case convert one form of energy (via energy converters - which were seen on-screen) into another one (such as using fusion power). 'Cause and Effect' showed both Warp nacelles as having power, and the Enterprise-D was hit while one of them was still active. To that effect - 'main power is down' could simply mean that while primary means of power is inaccessible, alternative power still (and does) actually run through the ship, or that main power is in insufficient quantities to power the ship, but enough remains (or that of the alternative source) that would be able to cause a catastrophic failure that can backtrack to the Warp core and cause an overload. The nacelles are one of the main components of Warp drive after all, and the main ship is inevitably connected to the main Warp core as primary means of power (as it would appear) - so even if alternative power is used, it could still use the same method of delivery which also allows going very near the Warp core or through it.
It's probably noteworthy that we did once witness a true shutdown where the nacelles went dark - in "Skin of Evil". If "some power remaining" is a key factor in disasters where nacelle damage propagates to the warp core, then "Cause and Effect" would probably fit those parameters while the ship in "Skin of Evil" would have been in no danger of such propagation. Since we have never heard of a starship powering down the warp systems for (anticipated impulse) combat, nor have we seen the nacelles go dark for such an occasion, we can decide that this is not a) a surefire way to protect the ship, or b) a practical way to protect her. That is, either Picard might make his ship a bit safer by shutting down all the glow in the nacelles but not so much that he should bother, or then it would be very, very difficult to shut down the glow. "Skin of Evil" emphasizes the difficulty of restarting after dilithium swap, but otherwise doesn't inform us about the possible difficulty of shutdown or restart. So I personally choose to belive a). Which means that the E-D blowing up was more a freak accident than the inevitable result of nacelle damage. Although whether the Bozeman also blew up or not, we can't really tell, because the camera only followed the E-D. Timo Saloniemi
Even before the Odyssey was finished off by the suicide run it had taken a horrendous beating. With more than one nacelle hit and a whole string of hits to the engineering section form the Jem'hadar attack ships.
The Nacelle like anything else can no doubt sustain a certain amount of damage before catastrophic failure. It is not beyond the realms of possibility, that given enough time plasma flow to the nacelle could be stopped to prevent it exploding or perhaps there is a mechanisim that they can eject the nacelle from the pylon. In the case of the Odyssey and Ent-D (in "Cause and Effect") there waswn't enough time for these failsafes to kick in.