Apologies if this has been posted before, but I couldn't find it here, and have only just seen the paper myself. It's by Julia Newton and colleagues and was submitted by them in 2012. There was no control group and they used the Fukuda case definition criteria. Interesting though. Full text here.