Well, she's wrong about the deletion being common among ME patient
I don't know what else she's claiming about it, but she usually overstates the significance of very common alleles rather massively.
The research I'm looking at indicates a lot of variation in which version is protective based on ethnicity or arbitrary sub-grouping after the data was collected. I'm not seeing strong results in either direction, just null, small, or non-significant variations in small patient groups, and nothing significant when directly measuring ACE enzyme levels resulting from the different genotypes. One cached article I couldn't directly access claimed a bigger difference, but I couldn't see any citations or basis for the claims. And the ones directly measuring ACE levels, such an
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2776311/ , show no variation in ACE enzyme activity.
One study did mention haplotypes being significant, in the context of bigger problems occurring when two highly linked SNPs aren't linked for someone (such as being homozygous for one SNP, and heterozygous or the opposite homozygous for the other SNP). But the 12 ME patients I have data for are completely linked in both SNPs.
So based on the research that I've looked at, rs4343 probably has little or no impact on anything, but might be interacting with other factors in some situations to create a problem. But there's not much indication what those other factor might be, aside from one or two haplotypes. Also, even when a common homozygous mutation does have a significant effect, the effect of a heterozygous mutation is usually small or non-existent, with a few exceptions.
If anyone can find a real research paper with better data, I'd love to see it, especially if it's looking directly at ACE activity rather than disease risk.