> Among people who smoke, the effect appeared to be synergistic, with B6 usage increasing risk threefold. The risk was even worse among smokers taking B12. Using more than 55 micrograms daily appeared to almost quadruple lung-cancer risk.
> There was no apparent risk among women—which is not to say it doesn’t exist, only that it wasn’t apparent.
Which would appear to rule out smoking as a confounder, except until you realise that people who smoke more might be more likely to take vitamin B6 - maybe because they are actually diagnosed deficient and not just trying to make themselves feel better - and that maybe B12 deficiency might be associated with higher levels of smoking as compared to lower levels (if higher levels of smoking are required for B12 deficiency than B6 , and the deficiencies are similarly diagnosed and not just feel-good vitamin supplementation). i.e. "smoking" isn't a binary category.
And the gender thing would be really surprising if the cancer was only caused by smoking - unless we know that smoking doesn't cause B6 or B12 deficiency in women, then that would also appear to remove smoking as a confounder.
In conclusion, maybe.