The article makes no mention of repeated measurements or a time period over which the subjects were studied. There's evidence of an association, but I see no evidence that exposure to air pollution causes diabetes. It might have been more effective had they studied these people longitudinally.

One confounding factor is that homes which are in direct proximity to high-traffic areas are cheaper than homes which aren't. The price factors in the noise pollution from the road, which also likely correlates with air pollution. It would be interesting to see what the relationship is between pre-diabetes and income, because I think that that's actually the relationship they're seeing.

We do know that poor air quality has detrimental impacts on the body, though, so I could be wrong.