It was exciting to read the paper by Liu and colleagues, one for which I had been waiting. When researching the literature for the meta-analysis I wrote on PD and non-melanoma cancers (showing reduced cancer risk ratios with relative risk of 0.73 and a 95 percent confidence interval of 0.63-0.83; referenced by Liu et al.), I came across many studies that individually showed an increased risk of melanoma in PD patients. By summing the studies available on the subject, the work by Liu et al. affirms the idea that an association between melanoma and PD exists. It is unclear why there is a reduced risk of some cancers (e.g., lung cancer, leukemias and lymphomas, colorectal cancer) and an increased risk of melanoma, and how these cancers are linked to PD. Certainly, future studies attempting to delineate the biological mechanism underlying these questions are warranted.

View all comments by Archna Bajaj

I would note that since the overall prevalence of melanoma in Parkinson's disease (PD) is pretty low, it is difficult to see a large role for melanoma-related mechanisms in PD, or vice versa.

Also, the extra medical attention that a subject gets after being diagnosed with a disease like PD may lead to increased detection of melanoma.

Finally, the PD cases in these clinical epidemiological studies were not verified neuropathologically, so there is an approximate 15-35 percent error rate built into the classification of PD versus non-PD (see Hughes et al., 2001 for clinical diagnostic accuracy rate).

In our Banner Sun Health Research Institute Brain and Body Donation Program, we have 121 neuropathologically verified cases of PD, of which three had a history of melanoma. Of the 750 subjects without neuropathologically verified PD, 23 subjects had melanoma. This comes out to a rate of 2.36 cases of melanoma to 100 PD subjects and 3.07 cases of melanoma per 100 cases of non-PD.

View all comments by Thomas Beach