Hi J.,
I ran a search of the National Institutes of Health library, the CDC site, and the FDA clinical trials site. I found NO documentation that any legitimate studies have been conducted on the metformin use as a causitive agent of congenital torticollis. If anyone claims that they've established a link, they're not basing it on sound medical/scientific research, and I would ignore it.
Doing a general search on metformin and birth defects, I found that one study in 2002 showed that there was a 30% decrease in the risk of birth defects in women with PCOS taking metformin, vs. those women with PCOS who were not taking metformin. In fact, the FDA has categorized metformin as grade "B", meaning that there are no risks of birth defects associated withit, although there is neither enough positive or negative information to draw a conclusion about its safety in pregnancy. In general, however, diabetic patients taking metformin are encouraged to continue usage during pregnancy as the risks to mother and baby from unregulated blood sugar levels is greater than any known risk of metformin usage.
Metformin works by helping glucose-resistant cells take up glucose so they can function properly. It was first developed for use by type II diabetics, but has now gained an off-lable usage for PCOS since the disease mechanisms are very similar.
BTW, I did notice an interesting clinical trial that will be recruiting soon, using botox and physical therapy to treat infants w/torticollis. If you're interested, check out
http://www.clinicaltrials.gov/ct2/show?term=congenital+to...