Can the Pupillary Light Reflex and Pupillary Unrest Be Used as Biomarkers of Parkinson's Disease? A Systematic Review and Meta-Analysis

Abstract

Background/Objectives: The pathological changes preceding the onset of Parkinson's disease (PD) commence several decades before motor symptoms manifest, offering a potential window for identifying objective biomarkers for early diagnosis and disease monitoring. Among the primary non-motor features of PD is autonomic dysfunction; however, its precise assessment remains challenging, limiting its viability as a reliable biomarker. Both the pupillary light reflex (PLR) and pupillary unrest are regulated by autonomic pathways suggesting their potential as objective non-invasive indicators of the PD prodromal phase. This review systematically evaluates studies that compare PLR and pupillary unrest in individuals with PD and healthy controls to determine their utility as potential biomarkers of the disease. Methods: A systematic search strategy was designed to identify studies reporting PLR and pupillary unrest findings in PD patients. Searches were conducted across three databases (MEDLINE, Embase PsycINFO), supplemented by cross-referencing relevant studies found on Google Scholar. The literature search was last updated on 7 December 2020. Pupillometric parameters that permitted statistical synthesis included maximum constriction velocity (VMax), constriction amplitude (CAmp), and constriction latency (CL). Pooled incidence and effect sizes were determined using a random-effects model with an inverse variance DerSimonian-Laird estimator. The I² statistic was used to assess study heterogeneity. When meta-analysis was not feasible, a qualitative analysis was undertaken. Results: The initial search yielded 219 references. Following deduplication and exclusion of ineligible studies, 31 papers were selected for review. Pupillometric data from 11 studies were incorporated into the meta-analysis. Effect sizes for PD patients were significant for VMax -0.92, (p < 0.01), CAmp -0.58, (p < 0.05), and CL 0.46, (p < 0.05). Measures of pupillary unrest were elevated in PD patients compared to controls, but evidence was limited to two studies. Conclusions: Pupillary constriction in response to light is characterised by reduced speed and amplitude in PD, with effect sizes suggesting potential clinical applicability. However, evidence regarding baseline pupillary variability remains insufficient, underlining the necessity for further research. Pupillary metrics represent a promising avenue for early PD detection, though their clinical utility is constrained by methodological heterogeneity and variations in disease duration among studies.

Keywords: Parkinson’s disease; biomarkers; neurodegenerative disease; pupillary light reflex; pupillary unrest.

Figure 1

Representative schematic of the response to a flash stimulus. The top panel illustrates the modulation of the parameter pupil diameter following the flash, highlighting the MCV (maximum constriction velocity) in the constriction phase and PDF (peak dilation velocity) during subsequent recovery. CAmp (constriction amplitude) is indicated on the left side of the figure. The bottom panel depicts the pupil diameter rate of change v(t), characterised by an initial deviation followed by stabilisation. The shaded region represents the duration of the flash stimulus.

Figure 2

PRISMA flow diagram detailing exclusions throughout each stage of study selection (PD = Parkinson’s disease) [62].

Figure 3

Newcastle–Ottawa Scale assessment of the studies included in the review. The percentage of studies awarded a point in each category is represented by an asterisk (*), while dash (-) represents studies not awarded a point.

Figure 4

Funnel plots of studies for (A) VMax, (B) CL, and (C) CAmp. Standardised mean difference from each study is shown on the x axis and standard error is shown on the y axis.

Figure 5

Forest plot demonstrating differences in VMax between PD patients and HC. The black boxes represent the SMD for each trial and the arms represent the 95% CI. The size of these boxes represents the relative weight of each study. The green diamond represents the overall SMD of the trials. * Study included twice, as two separate PD populations were tested.

Figure 6

Sensitivity analysis of all the included studies for VMax. Overall effect size shown as red dashed line with 95% CI shown as black dashed line.

Figure 7

Forest plot demonstrating differences in CL between PD patients and HC. The black boxes represent the SMD for each trial and the arms represent the 95% CI. The size of these boxes represents the relative weight of each study. The green diamond represents the overall SMD of the trials. * Study included twice, as two separate PD populations were tested.

Figure 8

Sensitivity analysis of included studies for CL. Overall effect size shown as red dashed line with 95% CI shown as black dashed line.

Figure 9

Forest plot demonstrating differences in CAmp between PD patients and HC. The black boxes represent the SMD for each trial and the arms represent the 95% CI. The size of these boxes represents the relative weight of each study. The green diamond represents the overall SMD of the trials. * Study included twice, as two separate PD populations were tested.

Figure 10

Sensitivity analysis of included studies for CAmp. Overall effect size shown as red dashed line with 95% CI shown as black dashed line.