On 2015 Feb 18, David Keller commented:

Should Parkinson patients with low LDL cholesterol increase dietary cholesterol intake?

Patients with Parkinson Disease (PD) tend to have lower LDL and total cholesterol levels than matched subjects without PD (and concomitantly lower rates of cardiovascular atherosclerotic diseases). The association of elevated cholesterol levels with reduced risk for PD incidence suggests that PD patients with abnormally low LDL and total cholesterol levels may benefit by elevating these lipid levels into the normal range, perhaps by consuming more dietary saturated fat or cholesterol itself. Of course, increasing LDL levels would also tend to increase cardiovascular risk; if increasing LDL by diet is found to slow down PD progression, then each PD patient will have to decide how much cardiovascular risk increase they are willing to accept in exchange for the concomitant neurological benefits. A randomized trial of atherogenic diet in PD patients would require careful attention to fully-informed consent, and should not raise the cardiovascular risk profile of any subject higher than the age-adjusted lower limit of normal for that patient.

On 2015 Feb 18, David Keller commented:

Evidence that high cholesterol, not statin use, is protective against Parkinson disease.

Prior studies have demonstrated decreased incidence of Parkinson disease (PD) with statin use, and with elevated LDL and total cholesterol. Subjects who are taking a statin are generally doing so because of a history of elevated LDL cholesterol. The reduced incidence of PD in these patients is difficult to interpret; is it due to their use of statins, or to their history of underlying elevated cholesterol levels? This uncertainty is due to "indication bias", so named because one factor of interest (elevated cholesterol) is an indication for the other factor of interest (a statin drug) to be prescribed. Prior studies which attributed a decreased risk for PD to statin use instead of the underlying elevated cholesterol may have suffered from indication bias.

Huang and colleagues addressed this problem by adjusting statin use for prior cholesterol levels, and adjusting cholesterol levels for current statin use. They found that high LDL and total cholesterol are associated with lower risk of incident Parkinson Disease (PD), and that statin use is actually a risk factor for PD. These findings imply that prior studies which reported a lower risk of PD with statin use may have suffered from indication bias, due to higher statin use among subjects with a past history of high cholesterol, and to lower current cholesterol levels among patients currently taking a statin. Huang's findings suggest that statin drugs are not candidates to slow or stop the progression of PD, based on the 56 PD cases analyzed.

Would it be possible to apply the statistical adjustments employed in this study to re-analyze the data used in prior studies, to reduce the effects of indication bias and determine whether prior reports of a protective effect of statin use were actually measuring the effects of the history of high cholesterol in these subjects?

On 2015 Feb 18, David Keller commented:

Evidence that high cholesterol, not statin use, is protective against Parkinson disease.

Prior studies have demonstrated decreased incidence of Parkinson disease (PD) with statin use, and with elevated LDL and total cholesterol. Subjects who are taking a statin are generally doing so because of a history of elevated LDL cholesterol. The reduced incidence of PD in these patients is difficult to interpret; is it due to their use of statins, or to their history of underlying elevated cholesterol levels? This uncertainty is due to "indication bias", so named because one factor of interest (elevated cholesterol) is an indication for the other factor of interest (a statin drug) to be prescribed. Prior studies which attributed a decreased risk for PD to statin use instead of the underlying elevated cholesterol may have suffered from indication bias.

Huang and colleagues addressed this problem by adjusting statin use for prior cholesterol levels, and adjusting cholesterol levels for current statin use. They found that high LDL and total cholesterol are associated with lower risk of incident Parkinson Disease (PD), and that statin use is actually a risk factor for PD. These findings imply that prior studies which reported a lower risk of PD with statin use may have suffered from indication bias, due to higher statin use among subjects with a past history of high cholesterol, and to lower current cholesterol levels among patients currently taking a statin. Huang's findings suggest that statin drugs are not candidates to slow or stop the progression of PD, based on the 56 PD cases analyzed.

Would it be possible to apply the statistical adjustments employed in this study to re-analyze the data used in prior studies, to reduce the effects of indication bias and determine whether prior reports of a protective effect of statin use were actually measuring the effects of the history of high cholesterol in these subjects?