A pesticide called heptachlor epoxide and used in 1970s was found in milk at that time - and it being linked to Parkinson's disease now, in a paper in Neurology.

It's not the first paper to link dairy products and Parkinson's disease, though correlation remain spurious - it is only in the last century that life expectancies have climbed high enough that people are not routinely dying from lots of other things before they get diseases of old age.  Academics have been scrambling to link thinks to various conditions and the authors say theirs is the first attempt to make milk causational.

To create the link, 449 Japanese-American men with an average age of 54 who participated in the Honolulu-Asia Aging Study were followed for more than 30 years and until death, after which autopsies were performed. Tests looked at whether participants had lost brain cells in the substantia nigra area of the brain, which occurs in Parkinson's disease and can start decades before any symptoms begin. Researchers also measured residue of heptachlor epoxide in 116 brains. The pesticide was found at very high levels in the milk supply in the early 1980s in Hawaii, where it was used in the pineapple industry to kill insects. 

The study found that nonsmokers who drank more than two cups of milk per day had 40 percent fewer brain cells in that area of the brain than people who drank less than two cups of milk per day. For those who were smokers at any point, there was no association between milk intake and loss of brain cells. Previous studies have shown that people who smoke have a lower risk of developing Parkinson's disease.

Residues of heptachlor epoxide were found in 90 percent of people who drank the most milk, compared to 63 percent of those who did not drink any milk. Abbott noted that the researchers do not have evidence that the milk participants drank contained heptachlor epoxide. He also stated that the study does not show that the pesticide or milk intake cause Parkinson's disease; it only shows an association.

There are numerous confounding factors. An obvious possible explanation is simple chance.
"Also, milk consumption was measured only once at the start of the study, and we have to assume that this measurement represented participants' dietary habits over time,"said Honglei Chen, MD, PhD, with the National Institute of Environmental Health Sciences at Shiga University, who wrote a corresponding editorial.