LAUSR

Visual dysfunction is increasingly recognized as a common nonmotor aspect of Parkinson’s disease (PD), with loss of visual acuity and color vision as well as higher order visual deficits. A related observation is that patients with PD with visual dysfunction may have a more aggressive cognitive phenotype with faster conversion to dementia. However, there are discrepancies between studies, with some showing no difference in visual acuity and higher order visual measures between patients with PD and age-matched controls. Differences are likely to arise due to heterogeneity in patients with PD, with sampling differences between small studies, many of which are cross sectional. For this reason, large longitudinal prospective studies of visual changes in PD are valuable resources. In this issue of Movement Disorders, 2 large epidemiological studies examine the link between visual dysfunction and PD. Both take advantage of established clinical databases with secure PD diagnoses and accurate measures of visual function captured in their relative populations. These studies highlight 2 distinct and clinically relevant aspects of visual dysfunction in PD: (1) that loss of visual function may be found at the very earliest stages of PD, a kind of early nonmotor sign, and (2) that visual dysfunction in established PD is a marker of poor prognostic outcomes. In the first study, Han and colleagues present a very large sample of more than 6 million people in South Korea with demographic and clinical measures as well as visual acuity, covering a study period of almost 9 years. This allowed them to identify measures associated with higher risk of developing PD. In addition to established risk factors such as nonsmoking, higher age, and diabetes, they observed that visual acuity was poorer in those patients diagnosed with PD during the observational period compared with those who did not develop PD (logMAR 0.41 ± 0.56 in patients with new PD vs. logMAR 0.13 ± 0.48 in those who did not develop PD; note that on a logMAR scale, higher values represent poorer vision). They also found that patients with poorer visual acuity showed significantly higher incidences of developing PD than those with good or normal acuities, with a hazard ratio (HR) of 1.36 and 1.27, respectively, for the 2 groups with poorest visual acuity (using best visual acuity patients as the reference and adjusting for age and sex). Interestingly, they did not find a dose effect. Although poorer visual acuity was linked with higher risk of PD, when they stratified by severity of visual acuity, the very poorest levels of visual acuity (with visual acuities poorer than 10/100) did not show the highest hazard ratios for developing PD, especially after adjusting for age, sex, and other covariates. The authors suggest that this may represent lower diagnosis rates in the group with poorest vision as a result of lower health-seeking behavior in this group. However, an alternative explanation is that there is a genuine inverted U-shaped curve here, with PD-related visual dysfunction producing a mild phenotype, and where severe visual loss is seen, causes other than PD are responsible. Indeed, where visual deficits are seen in studies, differences between PD and controls are not profound. It is our experience that visual dysfunction associated with PD is often asymptomatic and that patients with deficits on formal testing frequently are not aware of them in everyday life. An important potential confounding factor here is age, as increasing age is linked with both higher risk of PD and poorer visual acuity. This does not seem to be adjusted for in all analyses. Where HRs are adjusted for age, effects do lessen. For example, in the Cox regression analysis, when adjusted for age (and other demographic factors), effects are reduced from 6 times higher in groups with poorest visual acuity to 1.3 times © 2020 International Parkinson and Movement Disorder Society