True for pressurized water reactors (PWR), there's a primary cooling loop under enough pressure to not boil, and a secondary loop that boils to provide steam for the turbine.
But, there are a fair number of boiling water reactors (BWR) that generate steam in the reactor pressure vessel and pipe this straight to the turbine. The turbine shafts exit from the condenser side of the turbine, where the condenser provides a pretty strong vacuum. That causes air to be sucked in at the shaft seals. So, they have vacuum pumps that remove the air, to keep the condenser working properly.
Unfortunately, any radionuclides that leak into the cooling loop end up going through the vacuum pumps. Remember the Fukushima plants had those tall "smokestacks" painted with red/white stripes? Those were to vent off the gas from the turbines, and they had a steady stream of nuclides in them under normal operation.
They do water treatment on PWRs that removes the gases and nuclides, but in that case they can capture them in tanks and let them decay before release.
Jon
You still need two separate failures before isotopes from nuclear fission get released into the atmosphere:
-one of the fuel cells has to rupture/get a tear in order to get those isotopes into the cooling loop
-the containment of the cooling loop has to fail to get them into the atmosphere
Yeah, could be that one of the RBMKs in Leningrad has had a failure (this somehow sounds familiar…). If i remember correctly those designs didn't employ a second containment around the core, so they could have had a ruptured fuel cell, had to remove it from the core and it vented some gas when it was out of the reactor->not everything went through the filters. Makes you wonder how on earth they planned to safely remove broken fuel rods from this reactor design...
(Edit: that still wouldn't explain the clearly airborne caesium: Those reactors use light water as a coolant, the caesium would have reacted with it to form highly soluble caesium hydroxide->perfectly bound in the water of the cooling loop. So we can conclude that the source of this radiation spike was probably NOT sitting inside a water pool (water would have also captured dust too).)
The isotope distribution (Cs-134 vs 137) could give a hint of how old the damaged fuel cell was (extremely old would point to subs, newer fuel to a power plant), but i'm not sure if they are going to release that data to the general public. And even then, you'd need data at which point in the fuel cycle each isotope gets created, and i'm sure that such data can't be found on Wikipedia.