First a question: Are the antibodies detected with a secondary HRP-labeled antibody or by the inhibition of the MPO? So which enzyme converts the Isoluminol? Maybe the MPO is interfering with the marker enzyme?
Are you doing it in a white Mikrotiterplate? Sometimes the wells can "cross talk" if you don't use white plates. Also you should be aware, that you should do all incubation steps in the dark. These white plates somehow (Don't ask... its magic... I did not want to believe it myself the first time either!) the surrounding light and release it when they are in the dark detection device.
Have you tried to use your own washing buffer? Maybe PBS +Tween 20? I am sure you know, that Azide Ions inhibit Peroxidases, as well as CN, CO, a too high concentration of H2O2 and you your case maybe Chlorid-ions?
How did you notice the loss in sensitivity? Do you have a standard antibody, that you can spike in buffer?
Ok this is read-out is quite new to me. But is it possible, that the MPO reacts with the isolumniol and catalysis the oxidation before the signal read-out? So it "flashes" before it is supposed to be luminescent?
But maybe its not the read-out itself, maybe the secondary antibody is a little bit old and is sensitive to change in conditions?