It isn't the first time that some scientific experiments were not always reproduced in independent experiments.
One particular early experiment in electricity showed how a magnetically charged needle would move when put in a field caused by a coiled wire and have that needle change orientation. You would think this is a no-brainer and even something taught in junior high schools today.
Unfortunately this experiment was done by researchers who had their labs and lecture rooms oriented so the field was oriented north and south and didn't deflect the needle yet in other places it would work... simply because they weren't taking into consideration the Earth's magnetic field and that the direction of the experiment was thus an important factor.
I could give other examples, but sometimes experiments can't be repeated simply because not all of the variables have been addressed either by the original researcher or by the subsequent follow up experiments.
I could cite examples of Muon-catalyzed fusion (something that has ample experimental evidence) where some of the principles of "cold fusion" can indeed take place. Palladium crystals are also very famous for their absorption of Hydrogen atoms.
Based upon some physical experiments I've seen myself, there seems to be some actual fusion activity taking place in these crystals that could be called "cold fusion", at least in terms of detecting neutrons and some possible secondary detections of that activity. Those who think it could provide a practical source of power on the other hand are folks that I personally think are full of BS and just hyping things up for their own funding. You can even build a more practical "hot fusion" reactor that fits on your desk for a modest amount of money to produce even more verifiable reactions by simply building a Farnsworth-Hirsch Fusor, but I digress on that point. The main issue isn't if cold-fusion takes place (it does), but those who claim to have found a way to make it practical for power production.
You can have extremely local high density pressures where it is presumed that somehow two or more Hydrogen atoms are confined in a crystal matrix of multiple other atoms to create a state where fusion has a measurable probability of occurring. That at least is the theory behind how it works. The trick is to be able to open that door into that little pocket, just a few atoms at a time, and then close it up again to make the fusion take place. It could be Buckyballs, Palladium, or some other substance but it is some pretty wicked nanoengineering in order to get this to work. This is also why it is hard to replicate or even to make in the first place.
The problem with major physics journals is that so many frauds and perhaps just misguided fans have submitted papers that they simply don't have time to sift through them to find any real science on the topic any more.... other than perhaps minor tweaks of existing papers that were around before the Pons & Fleishman fiasco.