I've used and taught the use of various respiratory protection systems for a long time, including spending thousands of hours in training myself, teaching others in real fire conditions, along with emergency responses over 28 years. A single unprotected breath close to the actual fire is enough for permanent injury or even death. So I know a thing or two about the importance of getting a good seal.
It is true that anything other than bare skin where the mask touches the face can compromise its effectiveness. The question is more along the lines of how much compromise it causes, and how likely they are to ever encounter a need for their gas masks. Regular troop gas masks operate on a negative pressure principle. When someone breathes in, the pressure inside the mask is lower than atmospheric pressure, so any leaks are likely to allow inflow of contaminated air.
Firefighting mask on the other hand are positive pressure. They maintain a slightly above atmospheric pressure inside the masks, so theoretically a small likely will allow clean air out, but not allow contaminated air in. But, that pressure gradient either needs to be supplied by compressed air, or an external power source like a PAPR. That is basically a battery powered fan pushing air through the filter media to maintain pressure gradient. Not as easy or simple as a regular gas mask.
OSHA and most fire departments require anyone in an operational position to be clean shaven, but there has been both religious and medical challenges to that. But we also use computerized quantitative fit testing machines to ensure an appropriate seal.
With as much as it costs to train military folks, I'd tend to think fit testing them vs separating them would be a better option. Pull them occasionally like a random drug test so they don't do something different when they know a test is coming, and if they pass, leave them alone...